33[00:04:50] <ntz> what protocol is between camera and my linux ? I don't seee that in df -h ... I just have nikon camera connected using usb to my laptop, it works however the protocol / connection fails often and it's very slow
34[00:04:58] <RustyShackleford> blackflow: oh I missed your earlier comment. No discrete gpu
35[00:05:07] <RustyShackleford> would unplugging the monitor remove this variable?
56[00:08:05] <RustyShackleford> why does hardware hate me?
57[00:08:12] <RustyShackleford> ever electronic I buy develops gremlins
58[00:08:16] <ntz> jelly: hmm ... it's weird because (I use kde) if I try dolphin it just on normal file protoco; fai;s often but gwenview works relatively stable (importing large data)
59[00:08:17] <RustyShackleford> if it didn't come that way out of the box
60[00:08:20] <blackflow> ghost in the machine
61[00:08:21] <jelly> I've got servers with 9 year old drives that are dying!
63[00:08:51] <ntz> jelly: just nikon camera (d7200) connected using usb
64[00:09:03] <jelly> ntz, might be using one of those weird MTP, PTP or whatever protocols that have problematic linux implementations
65[00:09:18] <ntz> jelly: it uses some protocol that it's recognized but I have no idea what tools are managing that
66[00:09:24] <blackflow> jelly: and no ZFS on them to protect against data corruption? :)
67[00:09:27] <edi> hi guys, i'm experiencing something I would like your input.. I have a debian box on a lan that is downloading files extremely slowly, however the upload speed is normal, i tested it against a second box in the same lan. So I'm pretty sure it's a problem with the box itself, but I can't seem to figure out what. Iptables?
68[00:09:46] <jelly> blackflow, I don't have ECC RAM anyway.
69[00:09:57] <blackflow> doesn't matter much
70[00:10:02] <RustyShackleford> I could remove ALL variables, but the drive this came with in with Windows preinstalled
71[00:10:07] <RustyShackleford> and see if this gremlin goes away
72[00:10:11] <blackflow> the "zfs scrub of death" due to non ECC is myth and fud
73[00:10:24] *** Quits: noboruma (~noboruma@replaced-ip) (Remote host closed the connection)
74[00:10:59] <jelly> oh, one those servers, they're mostly just raid1 for OS
77[00:11:29] <jelly> zfs would not do a lot better given just two disks
78[00:11:49] <blackflow> what do you mean?
79[00:11:53] <ntz> jelly: thanks for input ... it's fine .. I don't have real issue ... I will keep googling ... it's indeed problematic because when I try to load a foldar that has a lot of data inside (? > 5GBs) it just fails however gwenview import works ...
80[00:12:10] <ntz> ^^ I mean fails when opening that as just usb device in dolphin
81[00:12:54] <ntz> I wanted to investigate how gwenview access that Vs dolphin
82[00:13:18] *** Quits: holden- (~holden-@replaced-ip) (Remote host closed the connection)
94[00:16:31] <Vamp898> In the Debian Reference Manual in Chapter 12.1 it says that btrfs will be, somewhen in the future, replace mdadm and LVM by combining both of its features. I want to state that btrfs does _not_ contain any logical volume management and there is no such feature on the roadmap
95[00:16:31] <jelly> blackflow, i mean what can zfs on 2 disks do better than md raid1 on two disks or even hw raid on two disks.
96[00:16:52] <Vamp898> So btrfs will neither now, nor (at least not planned) in the future be able to replace LVM
108[00:19:22] <Vamp898> whislock: many, yes. But LVM means "Logical Volume Management", and btrfs can't do that. It can't do Logical Volume Management, so it can't replace LVM
109[00:19:43] <whislock> Vamp898: One doesn't replace functionality with $exactSameThing.
111[00:19:54] <jelly> Vamp898, if you want the documentation to be fixed, can you please file a bug report on debian-reference package after checking if there isn't one there already
112[00:19:55] <RustyShackleford> if I can figure out a way to get more drives in this dell, I would set up something like zfs
113[00:20:14] <RustyShackleford> can I get an expansion card which will give me more sata connectors?
114[00:20:18] <whislock> Vamp898: LVM is a widget that satisfies use cases. If btrfs can satisfy those same use cases more effectively, then it is a replacement, even if it does $thing in a different way.
115[00:20:19] <RustyShackleford> there are only two on this motherboard
120[00:21:47] <Vamp898> blackflow: ? i dont get what you want to say. It behaves like a volume manager? I can not manage logical volumes with it. That is not that hard to understand or? It is not possible to create two raid6 and span one fs over them with btrfs, it just can't do that, you need LVM for this now and in the future.
138[00:25:26] <whislock> And the argument is sort of absurd, anyway. LVM can't do RAID60, either.
139[00:25:27] *** Quits: Envil (~envil@replaced-ip) (Remote host closed the connection)
140[00:25:28] <blackflow> now that you mention it... yeah btrfs works differently. it has no concept of vdevs (so it can't stripe them), it instead works with extents and can have different "raid level" extents on drives at the same time
141[00:25:39] <Vamp898> blackflow: btrfs cant, no. btrfs is a filesystem and if you create a RAID with btrfs, it still is an filesystem in the end. Adding more RAIDs would mean createing and filesystem on the filesystem and destroy everything (or better said, the btrfs tools dont allow to do that)
142[00:26:09] <Vamp898> If you have 24 Disks, and you want to make 2xRAID6 out of them, you have two filesystems with btrfs
143[00:26:13] <whislock> Vamp898: ZFS is also a filesystem, yet you're using that in this comparison.
144[00:26:20] <Vamp898> whislock: its not
145[00:26:24] *** Quits: morsicus (~morsicus@replaced-ip) (Remote host closed the connection)
146[00:26:36] <Vamp898> whislock: ZFS is also an filesystem, beside being an LVM
147[00:26:37] <blackflow> I think Vamp898 has a point, btrfs works quite differently than ZFS
149[00:26:57] <whislock> Vamp898: "Zettabyte File System". The hint's in the name. At this point, you're arguing things about which you are obviously misinformed solely to argue.
150[00:26:59] <blackflow> you can have both single and, say, mirror extents at the same time on the drives.
151[00:27:10] <whislock> blackflow: Yes, they do work differently. That isn't the point at argument, here.
153[00:27:29] <blackflow> whislock: I think it is because btrfs has no concept of vdevs
154[00:27:34] <whislock> The argument is that btrfs can't possibly ever satisfy the same use cases of LVM because it works differently, which is invalid.
155[00:27:40] <jelly> whislock, this is not a channel for arguments at all, we're here for tech support :-)
156[00:27:40] <Vamp898> whislock: Okay then let me say it different. ZFS can do LVM, btrfs can't
167[00:29:15] <whislock> Vamp898: A logical volume in ZFS is a dataset. Nothing more. Same as btrfs subvolumes. You're so hung up on names and terms that you're ignoring function.
170[00:29:49] <fakefur> this is not the right place to ask about debian testing is it?
171[00:30:00] <jelly> !debian-next
172[00:30:00] <fakefur> ah got it
173[00:30:00] <dpkg> #debian-next is the channel for testing/unstable support on the OFTC network (irc.oftc.net), *not* on freenode. If you get "Cannot join #debian-next (Channel is invite only)." it means you did not read it's on irc.oftc.net.
176[00:30:45] <Vamp898> whislock: btrfs subvolumes cant be used for logical volume management, so you still need LVM, that is my point. Its a bug in the Manual and does have to be fixed, the btrfs developers confirmed that to me a while ago so if they say btrfs can't do it, it can't do it
178[00:32:27] <blackflow> (but using lvm with btrfs would be so pointless since btrfs, as a pooled fs, is functionally doing the same)
179[00:33:11] <Vamp898> blackflow: you can't combine two RAIDs which is pretty common for big JBODs. There is no other way than using mdadm/lvm, there is no other option
181[00:33:30] <jelly> Vamp898, that's one thing LVM can do, but LVM does more things than literal management of volumes
182[00:34:01] <Vamp898> jelly: That is true, but the manual says btrfs will replace LVM, which is just wrong. It can replace parts of it, but it can never replaye LVM per se.
184[00:34:57] *** Quits: xcm (~xcm@replaced-ip) (Remote host closed the connection)
185[00:35:11] <jelly> that honestly sounds like something written way back when btrfs was new and was to be panacea
186[00:37:54] <blackflow> combining two raids under btrfs would be nonsensical. btrfs works best with individual drives so you give all those drives to it to manage.
187[00:38:15] *** Quits: towo` (~towo@replaced-ip) (Remote host closed the connection)
188[00:38:17] <Vamp898> blackflow: So you want to put 48 HDDs in one RAID6?
189[00:38:43] <Vamp898> blackflow: Out of 48 HDDs, the chance that more than 2 fail is pretty high, to high for me tbh
193[00:39:31] <Vamp898> blackflow: ? that is our actual setup. We have 4xRAID6 (12 HDD each), and those are combined into one lv using LVM and on top of that, there is a btrfs
200[00:40:39] <blackflow> which in no way changes the fact that given a saner number of drives (saner for btrfs), it would behave like a volume manager.
201[00:41:52] <Vamp898> jelly: none, the system is clustered + the whole data is backed up.
202[00:41:55] <blackflow> I'm not sure what btrfs does with that many drives for raid6, but I know for mirror, for example, it wouldn't do 48 copies. it'd do 2 copies and stripe all those drives, because it doesn't work with concepts like vdevs. you give it disks, and then you define raid level of blocks it puts on them.
204[00:42:47] <blackflow> logic would dictate it would do the same for raid6: one data block and two parity, spread among as many drives, but I'll admit I don't know that much about internals of btrfs.
205[00:42:48] <Vamp898> jelly: so the data exists 3 times, we are brave enough to run that without how spares
218[00:45:47] <Vamp898> jelly: We did, of course. Systems this big arent put into production like that, it had several stress tests. But the performance is mainly due to the 24 SSDs in front of the JBOD used for caching
220[00:46:33] <MrNaz> timedatectl status reports the incorrect timezone in raspbian current, but date does. if someone has a deb9 box they can check on, it might be worth doing and filing a bug report. i'd do it but i have no deb9 box up currently so i'm throwing it into the room here
221[00:46:37] <MrNaz> cheers
222[00:46:53] <Vamp898> blackflow: why would it be a wrong use case for btrfs? btrfs works pretty well and we need a very lot of the features ext4 can't offer
223[00:47:19] <jelly> Vamp898, honestly it sounds zfs would be a better fit
224[00:47:26] <Vamp898> jelly: ZFS can't do offline deduplication
235[00:49:13] <Vamp898> FinalX: out-of-band deduplication? Are we talking about the same thing?
236[00:49:16] <FinalX> every single time data is written it will cycle through your entire dedup table to find out if there's a matching block
237[00:49:27] <Vamp898> FinalX: That is in-band deduplication, we do not talk about the same thingt
238[00:49:50] <fakefur> does anybody know how i can disable the touchscreen on my laptop? it's a t440s
239[00:49:52] <fakefur> hanks
240[00:49:56] <fakefur> +t
241[00:49:57] <Vamp898> FinalX: out-of-band deduplication is only done when executed, in our case, during long weekends and vacation times
242[00:50:24] <Vamp898> FinalX: as long you dont execute it, it does nothing
243[00:50:35] <FinalX> yeah, I know, and yet it's still a very big table of data to hold, though it would indeed not be checked on writes
244[00:50:49] <blackflow> Vamp898: I think we have a slight case of misunderstanding here, but nevermind, like I said I don't know the internals of btrfs that well to know what it'd do with 48 drives and raid6 level blocks (but from what I see in the wiki, I'd assume it would stripe them achieving the same effect as having 4 x 12 raid6 arrays)
245[00:50:52] <FinalX> oh well
246[00:51:44] <fakefur> also i am unable to connect to oftc.net to verify my details to allow me to connect to #debian-next
247[00:51:58] <fakefur> does anybody know why that would be down?
248[00:51:59] <Vamp898> FinalX: With btrtfs the table can be stored on the File-System (in our case an dedicated SSD)
250[00:52:59] <Vamp898> blackflow: it just creates one big RAID6 containing all 48 disks, we tested it. We asked the btrfs developers in their IRC channel on how to change that and got the information, it is like that, btrfs can't do LVM
252[00:53:24] <Vamp898> blackflow: but its not stable anyway, so usind the btrfs own raid function is out of question for us
253[00:53:32] <jelly> fakefur, do you mean replaced-url
254[00:53:41] <fakefur> @jelly, yes
255[00:53:57] <blackflow> Vamp898: yeah this is purely theoretical discussion, btrfs can't even do raid6 with minimum number of drives properly, let alone 48 :)
256[00:54:42] <JackFrost> fakefur: Same nick there?
260[00:55:23] <joepublic> In defense of btrfs, it's the only mainstream-ish linux-kernel-supported filesystem that has live transparent compression.
261[00:55:28] <jelly> fakefur, seems to be down for me as well. I asked in #oftc channel over there.
262[00:56:05] <fakefur> @jelly, any other way to get some help with debian buster?
263[00:56:08] <jelly> I wouldn't trust a fs that can't figure how much free space is left to do compression
264[00:56:29] <joepublic> according to their wiki that's ok because it's "really hard"
265[00:56:41] <blackflow> ZFS doesn't seem to have that problem tho
266[00:57:28] <jelly> we get a btrfs user with disk full issues here about once a month
267[00:57:57] <joepublic> not saying btrfs is a good thing across the board, nor that other filesystems are not the awesomesauce.
268[00:57:58] <jelly> after ten years, it sure hasn't prompted me to go try btrfs for any use. let alone anything serious
269[00:57:59] <Vamp898> jelly: compsize gives pretty good results for us so far
270[00:58:05] *** Quits: Telcoguy (~quassel@replaced-ip) (Remote host closed the connection)
271[00:58:24] <joepublic> just, no other realtime compression filesystem outside abortive fuse things that do not, by design, work.
272[00:58:25] <annadane> "but the internet told me btrfs was good!"
273[00:58:48] <joepublic> annadane, I am sorry to be the one to say this, but things the internet tells us are not necessarily reliable
274[00:58:49] <blackflow> joepublic: uhm.... ZFS?
275[00:58:51] <fakefur> does anybody know how to disable a specific usb device?
276[00:58:52] <Vamp898> blackflow: The company i worked for before used FreeNAS as a storage backend for Citrix XEN, transparent compression was enabled and it had the exact same problem. Information about free disk space was more gambling and guessing than actually knowing
284[00:59:42] <Vamp898> blackflow: you get what i mean :D
285[00:59:54] <jelly> that's Linux' problem.
286[01:00:08] <jelly> use Debian on WSL :-)
287[01:00:11] <Vamp898> jelly: The company that used FreeNAS also uses Transparent compression on an 2008 R2 Windows server, but it was slow as hell, really really slow. that slow that copy operations aborted due to timeouts
288[01:00:19] <blackflow> Vamp898: no. ZFS will have issues with free space fragmentation but I have yet to hear of a situation that df says 50% is used and ZFS chokes up because it's really full to the brin.
289[01:00:19] <joepublic> and zfs is great, and zfs has compression, but zfs and linux are a source of controversy in a way that other filesystems are not.
302[01:04:30] <blackflow> joepublic: there really is no controversy. a few people shouting over politics with little regard to technical merit. meanwhile ubuntu is packaging zfs modules along with the kernel, and no Free/Libre software org challenged that legally.
303[01:04:35] <Vamp898> blackflow: but we do not use debian, we run an Gentoo with Linux Kernel 4.19.x and btrfs-progs 4.19.x, so syntax and/or features may differ for debian stable. I just noticed the sentence in the debian manual, thats the reason why im here.
310[01:05:53] <joepublic> blackflow, assuming there's no controversy, let's say it's a market penetration issue. one can take a debian install medium, install debian on btrfs, edit fstab to include the word "compress", and bob's your uncle.
317[01:06:49] <JackFrost> fakefur: Sorry, I meant is that your same account on OFTC?
318[01:06:51] <Vamp898> dvs: 4.19 is current LTS
319[01:06:53] <blackflow> no, 4.19 is the LTS
320[01:07:29] <dvs> No chance that a later kernel will be an LTS one before buster is released?
321[01:07:35] <a417> why?
322[01:07:37] <joepublic> even if you're totally right, which by the way, you are
323[01:07:38] <blackflow> joepublic: the only reason for that is that nobody sat and wrote the zfs functionality into the installers. meanwhile I've been debootstrapping debian with root on zfs on luks for quite some time just fine :)
324[01:07:42] <a417> 4.19 is plenty current
325[01:07:47] <Vamp898> dvs: 4.19 just came out, 4.20 is mainline
329[01:08:35] <a417> dvs: if it's that important to be newer than 4.19, just roll your own, it's actually gotten quite easy now
330[01:08:37] <blackflow> joepublic: which really ground my gears wrt ubuntu. they said they'd support zfs as first class citizen (and they package it up like that) but.... the installer still has no support for it, and yet they're now gonan rewrite the installer in electron. so..... priorities in NIH for that one. :)
365[01:19:41] <Vamp898> jelly: actually there is a difference between stable and longterm. stable kernels get a release almost every week, longterm kernels only get a release from time to time
366[01:20:12] <Vamp898> jelly: thats why we are already at 4.19.13
367[01:20:15] <joepublic> wouldn't that indicate that the less-frequently-updated $ITEMs were more "stable"
368[01:20:59] <Vamp898> joepublic: maybe, stable means nothing more than "out of mainline" and so "out of active development regarding new features and stuff"
388[01:25:34] <jelly> would be able to get to Linux 31.31 with just fingers
389[01:27:17] <Vamp898> jelly: i think kernel development just got way to fast. I mean one kernel release per week is already tiresome. Its not much more than "make oldconfig; make all -j32; make modules_install; make install; grub-mkconfig -o /boot/grub/grub.cfg" but im not used to such fast kernel releases honestly
392[01:29:36] <jelly> I stopped building kernels around Debian 5 or 6.
393[01:29:48] <Vamp898> jelly: i am looking forward for 4.19.13 to reach "real LTS" with slow releases every few weeks :D we even invented a new "micro maintanence" system to be able to kexec new kernels without announcments after 16:00
394[01:30:11] <jelly> that's cheating.
395[01:32:10] <Vamp898> jelly: it is. For VMs it doesn't really matter but for hardware machines thats a big performance increase in rebooting time :P
397[01:32:48] <jelly> just because silly UEFI firmware takes 8 minutes to POST
398[01:33:35] <Vamp898> jelly: we have some older hp proliant systems which take like ages to reboot. You could re-install the whole os faster in a chroot than the reboot takes
402[01:35:16] <Vamp898> Best thing is when team member change the hardware config like the ram, system complains about change memory and he just _temporarily_ accepts the changed memory
414[01:42:26] <Vamp898> oh wait, i think those servers might be Dell servers and not HP, the ones complaining about changed memory. I just catched that thought on my way to the toilet...
416[01:43:55] *** Quits: i1nfusion (~i1nfusion@replaced-ip) (Remote host closed the connection)
417[01:43:58] <blackflow> Vamp898: I beg to differ about LTS getting a release from time to time. On my gentoo box I see a new 4.14.x every several days.
421[01:45:36] <Vamp898> blackflow: we are on 4.14.91, thats 91 releases since 2017-11-12
422[01:46:17] <Vamp898> blackflow: i mean not we as a company, we as linux users :D
423[01:47:34] <Vamp898> blackflow: but i guess you're right. Until the development of 4.19 will slow down, it will already be EOL most likely. But we are on the latest LTS anyway, so it doesn't really matter that much. As soon the next LTS is out, 4.19 is dead for us
428[01:51:06] <Vamp898> blackflow: actually its weird how 3.16 had 62 releases yet and 4.4 had 169 already. I had the feeling 4.4 did progress slower... did it maybe increased in speed after its release?
433[01:55:14] *** jelly-home is now known as jelly
434[01:55:28] <jelly> 4.4 is managed by canonical people, they throw releases out as soon as possible. 3.16 is done by (debian's) benh, he aggregates lots of fixes and does not do point releases every week
435[01:55:31] <jelly> also, 4.4 is used by at least some android releases
436[01:56:50] <jelly> and now that this vm is rebooted successfully (with 4.20.0-trunk-amd64) I can sleep
437[01:57:49] <Vamp898> jelly: ah right, yes, but so is 4.9... i think we just have to get used to faster kernel releases
438[01:58:06] <Vamp898> At least for gk-h kernels
439[01:58:21] <Vamp898> jelly: n8
440[01:58:56] <RustyShackleford> are there nonfree drivers for intel graphics?
448[02:01:18] <RustyShackleford> trying to track down an intermittent fail to boot
449[02:01:26] <RustyShackleford> let me pastebin some logs
450[02:01:33] <rwp> Intel bought a company with non-free graphics so now there are actually *two* different Intel graphics. One with a free driver and one without.
463[02:04:17] <rwp> Was it PowerVR graphics? I am looking it up...
464[02:04:41] <blackbart> I have an asus usb wifi dongle; I installed the driver years ago and don't remember which one I'm using. Anyone know how I can figure that out? (I might have installed it from an apt package or from a github repo.)
465[02:04:46] <Vamp898> RustyShackleford: duckduckgo showed me this, there are two links there replaced-url
466[02:04:52] <joepublic> bad edid can come from a bad cable, an adapter such as displayport-to-something or something-to-something-else-like-vga-and-hdmi
467[02:05:48] <joepublic> sometimes adding a kernel option along the lines of drm_kms_helper.edid_firmware=OUTPUTNAME:edid/1024x768.bin can help get past it
479[02:08:09] <RustyShackleford> to me it seems like the system does not boot after the bootloader. Or at least it fails before anything can be written to logs
480[02:08:26] <RustyShackleford> I also ran a memory test and a drive test, all of which passed. I'm stumped
482[02:08:30] <blackflow> Vamp898: I find LTS releases more stable. By the time mainline get rid of quirks they're near EOL anyway. Just look at 4.19 and that whole ordeal with blk-mq corruption
483[02:08:49] <joepublic> when you step back and try to boot things like System Rescue CD (gentoo based), do you have the same problem?
486[02:09:33] <RustyShackleford> I could make some bootable usb and see if I can reproduce
487[02:09:39] <blackflow> Vamp898: on the other hand, LTS apparently tend to get less backported fixes. GKH and friends want us all to run mainline anyway and that's just too bad.
489[02:09:43] *** BrianG61UK_ is now known as BrianG61UK
490[02:10:27] *** Quits: MenschZwoNull (~MenschZwo@replaced-ip) (Remote host closed the connection)
491[02:10:47] <ryouma> a cve was released for jessie saying that vlc is vulnerable on mp4 videos. is there a solution, like backports, that allows me to keep running jessie and vlc, or do i have to upgrade just to keep vlc?
492[02:11:11] <ryouma> apparently the fixed vlc requires libraries not in jessie
495[02:12:18] <rwp> joepublic, Looks like the Intel GMA 500 is the problematic one. That's what is in my problematic system. There is a gma500_gfx driver. But it won't run X.
509[02:15:44] <rwp> Personally I think chromium should be in the stable-updates (aka volatile) suite and not in main. Since the upstream doesn't support a stable release.
510[02:15:55] *** Quits: alexandros_tab (~quassel@replaced-ip) (Disconnected by services)
524[02:18:23] <rwp> ryouma, At least the message is very clear and transparent about the problem.
525[02:18:35] <ryouma> yeah i like the message too
526[02:19:18] <rwp> Right, tzdata and similar. Things that change out-of-band to stable releases. tzdata in the US is "by act of congress" and they don't coordinate.
527[02:19:53] <rwp> Are you using the vlc GUI? Or just the command line interface? Because I normally use 'mpv' which works very well.
528[02:20:13] <ryouma> the gui. i think i ahve issues with mpv. like, no sound.
529[02:20:46] <rwp> Odd. Sometimes I need to specify a --lang=en to mpv to hint it to pick the correct language for me.
530[02:21:12] <ryouma> i did purge pulseaudio
531[02:21:17] <rwp> Also '#' will spin through all available audio tracks. If it isn't started with the right one then I bang on the # until I hear the one I want.
532[02:21:19] <ryouma> but vlc works great
533[02:21:50] <ryouma> huh. is there a mouse control for that?
535[02:21:57] <rwp> mpv is a fork of mplayer btw...
536[02:22:24] <rwp> Mouse control? Not that I know of. There is also a GUI for mplayer and I think also one for mpv too but as I said I don't use it.
537[02:23:50] <ryouma> huh, maybe i need to try to set associations for it. pcmanfm and deluge are the places i'd need it. do they call some debian wrapper?
556[02:33:03] <dvs> Oh that's not too bad. I hardly ran into problems myself.
557[02:33:08] <ryouma> i fear stuff not working, stuff going wrong etc. last upgrade broke x. i only rely on a few things, but thosethings are dealbreakers.
558[02:33:48] <dvs> ryouma, oh yeah. The network naming changed when I upgraded. Watch out of that.
617[03:07:24] <rwp> On server systems my upgrades from Jessie to Stretch were not eventful. On desktops I have had problems with X starting after the upgrade. But of course can work through it.
622[03:08:49] <rwp> Also if others do what I do on media pc systems of logging in remotely to play audio and video then I had to revert a bunch of stuff to allow that to continue working. Many regressions there. Unfortunately by explicit design.
629[03:15:41] <ryouma> hwinfo says "On Board Devices: #2560 Video: "Intel Graphics Media Accelerator 950" (disabled)" -- does that mean i can enalbe intel if i need to? it also says " Model: "ATI RV610 video device [Radeon HD 2400 PRO]"" which is ... amd?
630[03:16:24] <ryouma> rwp: hmm so i will need to be cautious
637[03:19:13] <rwp> I'm not positive but I think the Intel GMA 950 is a variant of the GMA 500 series. Which last time I looked was non-free. :-( (Times change however and my info might be old.)
638[03:19:30] <rwp> The Radeon HD 2400 PRO I think is covered by the free radeon driver and is okay.
639[03:20:35] <joepublic> there are "open source" radeon drivers but no "free" one afaik
644[03:22:11] *** Quits: poeticrpm (~poeticrpm@replaced-ip) (Remote host closed the connection)
645[03:23:01] <rwp> I caution against the proprietary amd/ati/firegl drivers as I have seen terrible stability problems with them. They seem to crash more than they work.
649[03:25:24] <ryouma> it turned out that x broke upon upgrade from jessie-1 to jessie, but installing microcode for radeon made it work again. modesetting issue.
650[03:25:27] <Vamp898> joepublic: the driver is in the linux kernel, so i'd say its free. Only the firmware is non-free
651[03:25:46] <ryouma> it took weeks to get that ifxed
652[03:26:04] <joepublic> "presence in the linux kernel" does not make something free, which is why there is such a thing as linux-libre.
654[03:26:40] <Vamp898> joepublic: linux-libre removes firmware blobs, the raedon driver is C-Sourcecode
655[03:27:04] <n4dir> joepublic: not that i'd be much in the subject, but i vaguley remember debian got rid of the itches gnu.org had with the kernel debian provides. just chat, of course.
656[03:27:12] <ryouma> is it dfsg free? if so then should the iernel be in non-free?
659[03:28:00] <joepublic> Vamp898, as you know, the free driver in the free kernel in the free world is nonfunctional without the nonfree package that contains the nonfree firmware...
660[03:28:48] <joepublic> ryouma, those parts of the kernel that are dfsg-free are in main. parts that are not, are not. everything in main is dfsg-free.
662[03:29:03] <Vamp898> According to the r600.c in the kernel Source (which is copyrighted by AMD and RedHat btw, the r600.c is completely free in like "do whatever you want with it" free
663[03:29:05] <rwp> Personally I'm okay with distributable blobs. I see them in a different catagory than non-free licensed software. Blobs are as good as the hardware in your box. How many people etch back the package and micropobe on the chip?
664[03:29:07] <joepublic> with debian main, there is no amd card with 3d acceleration
665[03:29:13] <Vamp898> The only think they mention is to keep the copyright notice intact
666[03:29:36] <joepublic> ...regardless of people who deny the fact.
670[03:30:44] <rwp> I go out of my way to find hardware I can use that has all drivers in the mainstream kernel so that things just work and I am not forced to fish my own solution.
675[03:31:00] <ryouma> when binary blobs are no longer supported, your hardware becomes a paperweight. i will never get a video card taht requiers a binary blob. which is why i got concerned that an upgrade to jessie required microcode (or ... whatever it is) just to get same fucnionality.
677[03:31:36] <joepublic> Vamp898, that's nice, but that code is non-functional. does nothing. does not work. is not a driver but rather a nonworking, nonviable, problem to be debugged. only when you combine it with the non-free package containing some firmware does it start to be useful.
678[03:31:42] *** Quits: dvs (~hibbard@replaced-ip) (Remote host closed the connection)
679[03:31:44] <Vamp898> ryouma: which graphic card does not need an firmware blob?
680[03:31:54] <ryouma> Vamp898: mine did not until jessie
681[03:32:06] <ryouma> maybe i am setting my expectations too high
682[03:32:08] <rwp> What does that mean ryouma? blobs no longer supported? They come with the hardware. Never change them. Just load them.
683[03:32:31] <joepublic> I know you don't want to hear from me because what I am telling you is inconvenient, but for the record, I am currently using an Nvidia GTX 760 card without nonfree firmware.
684[03:32:38] <rwp> If your card required a blob in jessie then it also requird it before jessie too.
688[03:33:11] <Vamp898> joepublic: the nvidia also needs a firmware, every GPU needs one
689[03:33:24] <ryouma> rwp: long ago, i had a matrox g550. and it required a binary blob. then x upgraded when i upgraded to wheezy or whatever. then i had no x. because it did not work with teh binary blob. matrox had nothing new. it took me 3 months to downgade x. why i didn't reinstall, idk. but that is the story.
690[03:33:53] <joepublic> Vamp898, the problem with your statement is that you are confusing firmware with nonfree firmware vs. reverse-engineered free firmware. one's free, one's not.
692[03:34:09] <ryouma> rwp: no, it woked without the firmware or whatever. x devekloped some mode setting requirement that broke it in both mode setting modes.
693[03:34:14] <rwp> I recall also having a Matrox g550. Definitely had several Matrox cards. Those were great back in the day. I don't recall any problems such as that with them. But it has been years.
695[03:34:15] <joepublic> I am glad you are stirred up about this--people often go through a period like this on the way to becoming free software activists.
696[03:34:53] <ryouma> i keep my hardware longtime
698[03:35:05] <rwp> Yes. Like I never used to care about trailing whitespace in source code lines. Now I am a fervent destroyer of trailing whitespace!
699[03:35:05] <ryouma> i am running like a 12yo box now
700[03:35:11] *** blackes__ is now known as blackest_mamba
701[03:35:34] <ryouma> lol i have to set up whitespace.el one of these days
702[03:35:59] <joepublic> there are some nvidia and many but not all intel GPUs that are 3d accelerated in the free world (no non-free components). Zero AMDs.
714[03:38:08] <joepublic> doesn't provide any non-free firmware for any hardware from anyone (Trisquel),
715[03:38:24] <Vamp898> joepublic: than this distribution is the wrong tool for your usecase
716[03:38:26] <joepublic> and gets decent frame rates for light duty 3d things.
717[03:38:41] <joepublic> Uh, it works great, does everything I want...?
718[03:39:00] <Vamp898> joepublic: i thought you were having issues with firmware
719[03:39:02] <Vamp898> im confused now
720[03:39:19] <joepublic> Vamp898, no just pointing out that there's no free amd driver. the driver requires non-free to work.
721[03:39:32] <Vamp898> joepublic: That is not an technical issue
722[03:40:17] <joepublic> it's a technical issue, in that you need to add the words "contrib non-free" to your sources.list if you want your card to work. Make sense? You have to specify what you want to add, and if the non-free firmware that makes your card work is what you want to add, that's how.
723[03:40:40] <Vamp898> That is not an issue
724[03:40:44] <ryouma> then x changes its interface and your card does not work
725[03:40:52] <Vamp898> Thats like saying, the command find is to long, i only want to type fnd, that means i have to setup an alias, thats bad
727[03:41:02] <Vamp898> ryouma: that is a technical issue
728[03:41:17] <joepublic> by default, debian does not provide such policies. It's an issue for many people who install debian. I am not following your 'alias' example.
729[03:41:46] <awal1> if i want to download all what listed here replaced-url
730[03:42:02] <awal1> not a regular playlist so not sure
731[03:42:14] <Vamp898> awal1: thats not legal
732[03:42:16] <awal1> the link is just an example
733[03:42:24] <rwp> ryouma, I know you remember being burned by that but I don't remember hitting that problem when I was using Matrox cards. But I do remember blobs previously getting installed by default but then getting pulled out of the distribution due to advocates who wanted to de-blob things.
734[03:42:32] <joepublic> Oh, cripes, you are a cop *and* a free software expert
752[03:46:02] <rwp> Agreed. I don't recall any case of it either.
753[03:46:03] *** Quits: CyberManifest (~CyberMani@replaced-ip) (Remote host closed the connection)
754[03:46:15] <ryouma> Vamp898: my case was when i did not ahve the firsmware package installed, and everything worked, then x (or the kernel) upgraded, and then x did not work without installing the (non-free) firmware.
759[03:46:38] <rwp> awal1, I think your best tool would be youtube-dl and if it doesn't do it stock then I would look at the latest upstream available. Since the sites change interfaces often.
777[03:48:40] <Vamp898> Stevie-O: Yeah, thats EFI. There is no bootflag like with MBR, EFI just looks at every disk if there is something to boot
778[03:48:41] <Stevie-O> Even though it's in a stupid drive enclosure hooked up via USB, it's taking priority over the actual drive inside the machine
779[03:48:53] <rwp> Stevie-O, The first thing I try to do in that case is swap SATA cables so as to reverse the BIOS/EFI order. Then it usually boots from the other drive.
780[03:48:56] <Stevie-O> okay, how do I persuade it that there's nothing to boot
781[03:49:11] <Stevie-O> rwp: there's nothing to swap. The offending drive is inside a USB enclosure
782[03:49:31] <Vamp898> Stevie-O: normally you can configure that. There should be an option in the EFI to what devices it should boot in which order
783[03:49:37] <rwp> Ah... So it boots the right thing if it is unplugged. But plugged in then it boots from the wrong drive. Gotcha.
784[03:49:46] <Stevie-O> Correct
785[03:49:55] <Vamp898> But its an EFI issue, nothing you can do on the harddisk other than purging the EFI boot partition from the disk
786[03:50:07] <Stevie-O> okay, so I have to change the partition type on the EFI partition
787[03:50:14] <awal1> joepublic, yeah that is what youtube-dl tell; but it does that for any fownload failed
788[03:50:18] <Vamp898> Stevie-O: you really should change it in the EFI Configuration
789[03:50:29] <joepublic> awal1, I mean by that that it's not just you
790[03:50:30] <Vamp898> Stevie-O: formerly known as "BIOS"
791[03:50:38] <awal1> rwp, i guss yes, youtube prevents download from music.youtube
792[03:50:59] <rwp> Patience and all things come to they who wait. It might in the future.
794[03:51:38] <Vamp898> ryouma: if Debian does have a regression, thats a debian thing. So you have three choices: 1. Accept that debian might break again with AMD Cards and fix it from time to time 2. Buy Debian compatible Hardware. 3. Change the distribution
795[03:52:03] <Vamp898> awal1: or you just buy the music
796[03:52:19] <rwp> Stevie-O, I think if you toggle the boot flag on the efi partition on that disk then it will no longer be bootable.
797[03:52:28] <Stevie-O> rwp: I don't know how to toggle that
798[03:52:46] <awal1> well, google may be doing what apple did historically. all youtube videos are uploaded by "normal people"; now youtube will commercialize all that
799[03:52:57] <rwp> Something like: "parted /dev/sdc set 1 boot off" at a guess and I didn't try it.
802[03:53:24] <rwp> Right. parted works with GPT okay. Also sgdisk works nicely with GPT.
803[03:53:32] *** Quits: Exmix (~Exmix@replaced-ip) (Remote host closed the connection)
804[03:53:36] <Vamp898> rwp: I can't remember that GPT does have an boot flag
805[03:54:00] <ryouma> so basically this box could die at any time and it would be good to know what cpu video or video card to get that will maximize the probability that debian will work for a long time using it
806[03:54:01] *** debhelper sets mode: +l 1429
807[03:54:06] <Stevie-O> interesting
808[03:54:20] <Stevie-O> rwp: That changed the partition type from EFI System Partition to Microsoft Basic Data
809[03:54:29] <rwp> I'm remembering that GPT does have a bootable flag. (But darn I just powered down my EFI test machine and can't check it for a few more minutes.)
810[03:54:32] <awal1> "upload everythg you want to my server/site; when I have enough content I make it commercial/non-free :P"
811[03:55:10] <Stevie-O> but that raises another question, which may be related to my pain in the first place
812[03:55:16] <Vamp898> awal1: well. YouTube is a commercial product from a private, commercial company
813[03:55:19] <Vamp898> awal1: what did you expect?
814[03:55:20] <rwp> Stevie-O, As I recall EFI requires a bootable flag. If it isn't bootable then it can't be an ESP (EFI System Partition)
815[03:55:21] <Stevie-O> all of the drives are labeled "debian" in the BIOS
816[03:55:39] <Stevie-O> I wonder if that's why it's getting confused about which drive to boot from
817[03:55:51] <awal1> Vamp898, you right
818[03:55:51] <Vamp898> awal1: But YouTube Music (what you are trying to download from) is not YouTube. Thats a Music Platform like Spotify
821[03:56:21] <Stevie-O> So how can I change the name that the BIOS sees?
822[03:56:29] <awal1> i thought it is part of the regular youtube
823[03:56:32] <rwp> I have seen that too Stevie-O and it seems very confusing to me. But UEFI is a design by committee thing. Those never go well.
824[03:56:59] <Vamp898> awal1: Nope. It does have its own App, own infrastructure and so on
825[03:57:14] <awal1> ok, ok. thanks for clarifying
826[03:57:19] <awal1> nvm, so
827[03:57:57] <Vamp898> awal1: you can check if the artist is on bandcamp. You can buy music there DRM Free in all formats, even FLAC. And the money, normally, goes directly to the musican. Thats why you often have choose what you want to pay
828[03:58:35] <awal1> ok, Vamp898
829[03:58:48] <awal1> (Y)
830[03:59:39] <Vamp898> ryouma: well that is a complicated topic. If this system is not dedicated Linux compatible, than its not. There is no guarantee that it will run always with Debian
853[04:06:37] <annadane> TIL. the only other linux computers i was aware of previously was thinkpenguin
854[04:07:17] <ryouma> probably walmart or so. there are white box stores but they charge a few hundred dollars markup. and there's ncix which claims to take a spec and build it for 75usd and send it to you but odn't answer their email.
870[04:11:37] <RustyShackleford> well with all original hardware in this machine, original disk with windows 10 installed, the intermittent boot issue is gone
871[04:11:56] <RustyShackleford> I wonder if its a combination of the hardware and the software that comes with debian
872[04:12:17] <RustyShackleford> maybe I'll try ubuntu server instead. maybe newer software will make the difference
873[04:12:28] <Vamp898> ryouma: That is one advantage with FreeBSD (even though i am not a FreeBSD Fan). They have a complete list of all hardware they support
874[04:12:47] <awal1> n2dir, last I checked thinkpenguin website and an other one (i forgot name) an i3 /4 or 6 go ram was +1000$
875[04:13:07] <awal1> Vamp898, 1300 from thinkpenguin or where?
876[04:13:25] <Vamp898> awal1: Tuxedo Computers
877[04:13:27] <awal1> n4dir not n2dir :D
878[04:13:28] <n4dir> awal1: yeah, sure, i do believe you. Just saying that not everyone has the same needs.
879[04:13:41] <n4dir> n2dir is even better :-)
880[04:13:41] <Vamp898> awal1: The InfinityBook Pro 13 v3 (no longer on sale, its now the v4)
883[04:14:12] <annadane> the thing i like about these linux computers websites is that it's *easy to surf them*. so many computer manufacturer websites are shit and obscure
884[04:14:27] <awal1> Vamp898, I will check that site, ok
885[04:14:48] <Vamp898> annadane: yes, and the others are like "Why are sales dropping" :D
916[04:22:44] <Vamp898> But i am not an free software actiivist. I use free software whenever possible, but i do have my limits. My Smartphone for example runs AOSP, but i of course use an Binary blob for the camera and wifi
917[04:22:45] <Kremator> hey folks, is that debian project of "GNU/freeBSD" still alive?
918[04:22:54] <Vamp898> I use an non-free driver for my printer
927[04:24:25] <Kremator> This is a release in progress. It has been released with Debian 6.0 (Squeeze) as a technology preview and the first non-Linux port.
928[04:24:25] <Kremator> " ?????/
929[04:24:53] <n4dir> i sure had a much later release than 6.0 installed
930[04:24:54] <Vamp898> Kremator: well, according to the XFCE Homepage, XFCE also sounds pretty dead
931[04:25:00] <awal1> dpkg: freebsd
932[04:25:00] <dpkg> rumour has it, freebsd is NOT better than Debian, and it is not worse. It is just different. Admittedly one might be more interested in using FreeBSD as a server OS rather than as a desktop workstation. replaced-url
933[04:25:05] <annadane> xfce is active
934[04:25:18] <Kremator> Vamp898, why you have to be salty
935[04:25:25] *** Quits: blanka (bad576bf@replaced-ip) (Quit: Page closed)
936[04:25:27] <awal1> dpkg: kfreebsd
937[04:25:27] <dpkg> Debian GNU/kFreeBSD is a port that consists of GNU userland using the GNU C library on top of FreeBSD's kernel, coupled with the regular Debian package set. kfreebsd-amd64 and kfreebsd-i386 are not official release architectures since Debian 8 "Jessie" due to quality concerns. replaced-url
938[04:25:45] <Vamp898> annadane: last release was in 2015, GTK+3 port is, with a lot of luck, finished shortly before the release of GTK+4
954[04:27:51] <Stevie-O> I wish to be able to upload files to that partition, but I *don't* want to enable root login via SSH, and I *don't* want my regular account to have free reign of the partition
955[04:28:06] <Stevie-O> suggestions?
956[04:28:21] <Kremator> you dont want root there but also you dont want any users there
957[04:28:22] <Kremator> check
958[04:28:29] <Vamp898> annadane: according to their roadmap, there are a lot of things that urgently need to be changed, so i'd say they are in lack of people who want to work on it
959[04:28:33] <Kremator> so, basically how anybody would ever do anything inside that partition
960[04:28:33] *** Quits: dvs (~hibbard@replaced-ip) (Remote host closed the connection)
961[04:28:50] <annadane> Vamp898, fair enough
962[04:28:52] <Stevie-O> well, in a shell I can sudo
963[04:29:00] <Stevie-O> but that doesn't really work for sftp
968[04:29:26] <watchcat> Stevie-O: ssh with root login diabled.
969[04:29:43] <Kremator> what about SCP?
970[04:29:58] <rwp> When you say "upload" how do you mean? A web server allowing uploaded files? Then set up a uid such as upload-data and only allow it access.
978[04:31:27] <Stevie-O> I didn't say it couldn't be my user account
979[04:31:41] <rwp> "and I *don't* want my regular account to have free reign"
980[04:31:43] <Stevie-O> I just said I didn't want my regular account to have free reign
981[04:32:04] <rwp> What's the difference between write access and free reign?
982[04:32:10] <MACscr> im trying to troubleshoot a system and happened to have a debian usb thumb drive setup to do net installs. I have dhcp enabled and got to its console. Its there a way to enable openssh from there? im just in the installer and hit alt + f2 to get the console
983[04:32:24] <Stevie-O> rwp: nothing, really
984[04:32:29] <rwp> But in any case create either a user or a group, allow that user or group to write to directories there.
990[04:33:14] <rwp> Since you want it locked down a user seems to match. But most people use shared work groups with the sgid bit set on the directory. chmod g+ws on the shared directory and chgrp $sharedgroup
1005[04:35:34] <MACscr> im just trying to do some stuff with the hardware
1006[04:35:40] <rwp> MACscr, Ooohhh! Well that changes everything.
1007[04:35:53] <rwp> You don't happen to have a live-boot system handy?
1008[04:36:04] <watchcat> as long as your regular account can do su or sudo, you can't restrict what it can do. that's why you need to make another account.
1009[04:36:37] <rwp> MACscr, Let me look for something...
1017[04:38:26] <Stevie-O> I can make a subdirectory that's writable by my account, and just move the files over from there
1018[04:38:30] <Vamp898> n4dir: Actually MAC (Mandatory Access Control) like Tomoyo, SELinux or AppArmor are exactly for that case
1019[04:38:32] <Stevie-O> what the frig
1020[04:39:18] *** Quits: ryouma (~user@replaced-ip) (Remote host closed the connection)
1021[04:39:33] <Vamp898> n4dir: sudo is an additional layer of security, but its pretty limited, especially if you are on something like a VM Host where several people easily can work as root
1022[04:39:36] <rwp> MACscr, Hmm... I was remembering this replaced-url
1036[04:42:31] *** Quits: JPT (~jpt@replaced-ip) (Remote host closed the connection)
1037[04:42:33] <Voldenet> when it comes to security I wouldn't blindly count on sudo though
1038[04:42:47] * rwp says Drat! I am short one network port on this switch. I'll be afk for a bit while I scrounge.
1039[04:43:04] <Stevie-O> rwp: did you see the thing where someone posted -- as a joke -- a "run this" that granted the author remote root access, and tons of people ran it?
1040[04:43:05] <rwp> I think the path you are traveling will eventually land you talking about SELinux restrictions.
1041[04:43:12] <watchcat> what you can do is require wheel for access to su, and make a user who isn't in either wheel or sudo group.
1042[04:43:16] <Stevie-O> like a "curl foo | sudo bash"
1057[04:48:42] <MACscr> if i do take a bootable image like a raspi image file and dd to that to a sdcard located at /dev/sdX, if that image is bootable, that sdcard should now be bootable, right?
1072[04:55:48] <MACscr> ive been having an awful time trying to get retropie to boot on my pi. always getting a vfs kernel error. Tried burning it from my mac and my ubuntu media server. same result. was using an old sd card reader. Figured maybe the reader could be a problem and/or the sd card. So now im trying it from an old computer i have that has a built in card reader and had to boot it up with the debian installer to get to a cli to try it again with
1078[04:57:01] <awal1> radio frequency, power consumption, hacked hardware and such stuff can be real :P
1079[04:57:05] <rwp> I *have* had problems with poor quality SD card adapters. And then switched adapters and everything worked perfectly. I don't trust them now.
1087[04:59:56] <annadane> and stole all of saddam's cat pictures. how tragic.
1088[05:00:09] <MACscr> rwp those are just methods for sending data though. you still have to infect the system in the first place and you might as well steal the data then
1089[05:00:14] <awal1> :D
1090[05:00:41] <awal1> now we have more democracy and stability in Middle east, btw :P
1091[05:01:07] <rwp> I was just reacting to your supposition that computers not connected to the Internet can't always be safe. :-)
1092[05:01:38] <rwp> I think that if someone has a C&C mill controller and it is not connected to the Internet then I am not worried about it.
1093[05:01:42] <MACscr> they were safe, until someone got physical access to them to install the malware
1106[05:08:48] <MACscr> and damnit, dd didnt seem to work. not even getting an hdmi response this time. so frustrating when simple things dont work as they should
1110[05:11:01] <rwp> Good point. It would need to be embedded in too much concrete to move too. (chuckle)
1111[05:11:33] <annadane> the concrete has a backdoor.
1112[05:11:39] <rwp> MACscr, If you suspect a problem writing the SD card then even though it takes a long time I suggest reading back what you wrote to verify it.
1141[05:41:17] <altker128> Hey guys. Anyone here ever installed the 389-ds package? Curious if it installs cleanly, what kind of web site should you get when apache2 loads
1213[06:39:45] <rwp> Sorry but I won't be able to help. I am one of the non-systemd folks. And this is one of those reasons. You will need a systemd person to help you.
1214[06:40:25] <RustyShackleford> rwp: what do you use instead?
1215[06:40:33] <RustyShackleford> it could be the particular version of grub also
1216[06:40:54] <RustyShackleford> trying with debian 10. If newer packages make my problems go away, I'm happy
1217[06:41:01] <RustyShackleford> even if I don't know the cause
1289[07:53:06] <jdhaddon> Yes. I'm not medically cleared to work at the moment, so I was gifted root access to a server running Debian 10. It has apache2 and nginx installed. I was asked to setup four websites; three blogs one forum.
1290[07:54:27] <jdhaddon> Essentially I have the free time and two friends have website ideas.
1293[07:58:53] <jdhaddon> One works as a developer and sysop, thats where we got the server access from. The other works for the railroad. So they both work 80+ hrs a week, no time to develop hobby side projects. And then theres me who has absolutely nothing but time right now
1298[07:59:59] <rwp> It's mostly personal preference but I much prefer working with Nginx over Apache. Apache is a huge memory pig by comparison. Nginx is light and very fast.
1299[08:00:17] <rwp> But Apache has been around for so long that most projects include Apache configuration by default.
1300[08:00:24] <jdhaddon> thats what I keep reading
1302[08:00:55] <rwp> Nginx is the new kid on the block. Therefore even though I find it much easier to work with that means one must configure it. But that is always my choice to do these days.
1303[08:01:03] *** Quits: kupi (uid212005@replaced-ip) (Quit: Connection closed for inactivity)
1304[08:01:46] <whislock> I use nginx as a proxy in front of Apache, personally.
1305[08:02:17] <rwp> Me too!
1306[08:03:01] <jdhaddon> Elaborate please!? Whats the benefit
1307[08:03:08] <rwp> Some things absolutely require Apache such as Subversion's WebDAV backend. The module has only been written for Apache. But I use Nginx as the frontend and proxy to Apache for it for example.
1308[08:04:06] <rwp> On one site that the client runs Wordpress and Apache it was only possible to run 18 Apache daemons before the system was out of memory. Running Nginx that could be cranked up to about 80. Just for a comparison.
1309[08:04:46] <rwp> Personally I just like the clarity of the Nginx configuration files. They are quite plain and say what they do and do what they say.
1310[08:05:14] <rwp> With Apache even though I have been working with it for a long time I always have to spend a lot of time with the docs open to figure out how things should configured.
1315[08:08:19] <jdhaddon> Could you break that down a little bit for me? Or point me in a direction that I can research things? 18 vs 80 daemons, whats that actually mean
1316[08:09:07] <rwp> The amount of memory consumed per process. Every daemon process servicing incoming requests uses some of it. A working set size of memory.
1317[08:09:35] <rwp> If you have a conference site and the organizers like to tweet things then it is like a coordinated attack on the web site.
1318[08:10:21] <rwp> They will tweet out something like, Updated schedule! Look now! or Rooms available, check availability NOW! and then 15,000 people will get the tweet and some 1,000 people will try to hit the site within seconds.
1319[08:10:39] <rwp> It's basically a coordinated attack, a DDOS like denial of service on the web site.
1320[08:10:58] <rwp> Therefore the site needs enough bandwidth in terms of networking and cpu resources to deal with the onslaught.
1321[08:11:04] <jdhaddon> I was just gonna say: so essentially a DDOS
1322[08:11:12] <rwp> Because the organizers don't understand that it would be nice to skew those somewhat.
1323[08:11:33] <rwp> To handle it one needs a lot of parallel processes running.
1324[08:12:32] <rwp> With Apache IIRC the process size for Wordpress as they had it configured was around 55M per process with about 1gig of ram available to run. That would only fit about 18 processes at 55M per before they had consumed all of the available ram.
1325[08:12:34] *** Quits: awal1 (~awal1@replaced-ip) (Remote host closed the connection)
1330[08:13:16] <rwp> Meanwhile Nginx doesn't process PHP directly. Nginx is just the web server. To server PHP one uses php-fpm (fast process manager) and another set of daemons for serving the PHP interpreter.
1331[08:14:01] <rwp> But the php + php-fpm combination is more efficient in terms of memory use. And it is faster. Therefore it can both support more parallel proceses and doesn't need as many at the same time.
1335[08:16:45] <RustyShackleford> what would you use?
1336[08:17:43] <rwp> I don't have any experience with node.js so can't say. But you can look and see how much memory does each process consume? Use ps such as "ps auxww" (bsd style) or "ps -ef" (sysv style) and look at the resident set size (rss).
1340[08:18:30] <rwp> Questions: Are you running a database on the same machine? How much memory does it consume?
1341[08:19:26] <rwp> How much file system buffer cache do you wish to maintain? Hint: zero means slow. I recommend as much file system buffer cache as you can. But it really depends upon what you are doing with the system. Enough to cache most working loads.
1342[08:19:57] <rwp> So after marking off file system buffer cache, and database memory, and a little bit of extra slosh, then the rest can be dedicated to the web server frontned.
1344[08:20:45] <rwp> Or put it another way... The Apache default number of processes is 256 and if each is 55M then it will cap out at 14 gig of ram just for the frontend.
1346[08:21:06] <rwp> If you had a 16 gig memory machine that would still fit. But if it were only an 8 gig ram machine then the machine would have gotten deep into swap thrash pretty quick.
1347[08:21:32] <rwp> When swapping the machine will be crawling in terms of speed. If it doesn't have the swap for it then the OOM Killer will be invoked.
1349[08:21:54] <rwp> It will start killing Apache processes. Hopefully. It might kill other random processes, perhaps very important ones. I hate the OOM Killer.
1350[08:22:24] <rwp> Therefore one must always restrict the number of daemon processes the supervisore can spawn to something that is below the amount of ram available on the system.
1351[08:23:01] <rwp> Hopefully all of that makes some sense.
1356[08:26:17] <RustyShackleford> yeah it makes some sense
1357[08:26:27] <RustyShackleford> the tldr is to use nginx?
1358[08:26:30] <rwp> It would make sense to you if you were running a production web server in that situation. :-)
1359[08:26:40] <rwp> Yes! Nginx FTW!
1360[08:27:28] <RustyShackleford> maybe i'll just use heroku lol
1361[08:27:49] *** Quits: morsicus (~morsicus@replaced-ip) (Remote host closed the connection)
1362[08:27:55] <jdhaddon> ok, so nginx seems like the likely choice. Apache is the current server, serving only one site. So how do I go about activating nginx as default, and then serving the four different websites? About configuring a proxy? I need to go walk my dog so I'll be away for 10-15 minutes, please excuse my absence
1363[08:28:05] <rwp> It all depends upon what you want. Commercial sites like Heroku have expertise for sale. You pay them and they serve your pages.
1383[08:30:50] <rwp> If you have a profitable idea and it takes off then you will have income to pay for the scale-out that you need to do to make it work.
1384[08:31:22] <rwp> If you are just hobbying around whichi is great, its how we learn things, then it won't be a problem. Just do one simple thing and have fun with it and learn from it! :-)
1393[08:36:06] <rwp> dx_ob, See the /usr/share/doc/i3blocks/README.md file for documentation on it. But basically it needs i3status changed to i3bar in the .i3/config file and then i3 restarted with $mod+Shift+r restart.
1423[08:47:47] <rwp> I used to use dchroot and then schroot but lately those have grown to be just too much and so now I have stopped using schroot and have gone back to just plain old chroot.
1424[08:48:21] <[rg]> rwp: really? I just used a script to set up schroot, was fantastic
1454[09:11:19] <jdhaddon> rwp: ok, I'm back. So where do I go from here? This is totally DIY, I just have access to better equipment than most. No commercial use, just hobbyists. What should I be researching to make nginx the primary server for four or possibly more websites in the future? How do I make nginx the primary, not apache?
1455[09:11:46] <rwp> The primary web server is the one on port 80 for http and 443 for https.
1456[09:12:24] <rwp> Both can be running as long as only one of them is on port 80 and only one on port 443.
1457[09:13:15] <rwp> What I would do is install both, edit /etc/apache2/ports.conf and change the ports there to another port that is unused, say 81.
1464[09:14:21] <rwp> At that point you will have both running. If you wish to proxy from nginx to apache then set up the configuration to do that. I'll prepare a paste...
1465[09:14:22] <jdhaddon> both are installed, but apache appears to be the default
1500[09:24:03] *** Quits: Thedarkb1-M90 (~Thedarkb3@replaced-ip) (Remote host closed the connection)
1501[09:24:07] <BenNZ> jdhaddon: so you want to make a web server ?
1502[09:24:12] <TommyKaira> im in long holidays
1503[09:24:27] <rwp> jdhaddon, Here is a simple complete example, just http port 80, that also proxies to a local (Apache) server on port 81: replaced-url
1510[09:27:15] <rwp> In the nginx config notice the server_name line. For every web site you wish to host there simply copy that config file to a new file, change the server_name to the additional name you are setting up.
1511[09:28:02] <rwp> And of course those go into /etc/nginx-sites/available/$SITENAME and get symlinked "ln -s ../sites-available/$SITENAME /etc/nginx/sites-enabled/"
1519[09:30:08] <rwp> I used that /cgit/ proxy example as an example where cgit is served by Apache. In your case it would be whatever it is in your case.
1520[09:30:19] *** Quits: milkt (~debian@replaced-ip) (Remote host closed the connection)
1531[09:36:19] <rwp> Then also must edit every :80 to be :81 in the /etc/apache/sites-available/* files. Really only the ones that are actively linked into /etc/apache/sites-enabled/* though.
1532[09:36:24] <jdhaddon> why 81 and not 80? Did I miss something?
1533[09:36:44] <rwp> You were saying you wanted Nginx to be the default, and wanted to learn how to proxy to Apache running on another port.
1534[09:36:45] *** Quits: cfoch (uid153227@replaced-ip) (Quit: Connection closed for inactivity)
1536[09:37:07] <rwp> For Nginx to be the default then Nginx must use port 80. Which means Apache must be configured to run on a different port.
1537[09:37:41] <rwp> Like "The Highlander" there can be only one process bound to a network port number.
1538[09:37:50] <jdhaddon> got it
1539[09:38:34] <rwp> Configuring a different port is simple enough and then Apache can continue to run without getting in the way of nginx that we want on port 80 http.
1540[09:38:41] <rwp> I'll ignore port 443 https for the moment.
1541[09:38:46] <rwp> Setting up Let's
1542[09:39:06] <rwp> Setting up Let's Encrypt for https TLS/SSL certificates is not that difficult. But more involved.
1559[09:43:38] <plasmoduck> hi, I need help mounting a 3tb usb hdd. when I do 'fdisk -l' it says '/dev/sda2' 'Type' 'Microsoft basic data'. So I think it's NTFS?
1560[09:43:39] <rwp> And if that is a misunderstanding then the Apache config can be reconfigured back to port 80 and no harm done. :-)
1561[09:43:42] <plasmoduck> But then when I do 'mount -t ntfs-3g /dev/sda /media/data' I get NTFS signature is missing.
1562[09:43:51] <plasmoduck> What am I missing?
1563[09:44:02] *** debhelper sets mode: +l 1436
1564[09:44:06] <rwp> You are missing the partition number.
1565[09:44:15] <rwp> Instead of mounting /dev/sda you would mount /dev/sda2
1599[09:52:02] <SerajewelKS> sda2 contains an MBR, which is unusual for a partition. so i asked him to run fdisk on that to see if it, for some reason, contained its own partition table
1600[09:52:09] <rwp> I am likely missing out on many details. The story of my life...
1601[09:52:29] <SerajewelKS> then i told him that he should use gpart, which is a tool that looks at the contents of the disk to try to guess where partitions _should_ be based on filesystem signatures
1602[09:52:44] <SerajewelKS> i'm gusesing the NTFS superblocks are offset because of some weird-ass microsoft way of doing UEFI
1603[09:53:04] <SerajewelKS> gpart would tell him where the filesystem really is, then he can losetup a device with an offset
1611[09:54:23] <SerajewelKS> gpart is not a partition table editor, it's a partition table _guesser_
1612[09:54:23] <rwp> jdhaddon, If you have gotten apache off of port 80 then you will be able to connect to it on port 81 (firewalls not withstanding that might prevent it).
1613[09:54:26] <jdhaddon> asking*
1614[09:54:40] <rwp> Then can start nginx and it will be on port 80 and be the default web server.
1615[09:54:55] <rwp> And can then proceed to use it from there on out.
1616[09:55:24] <rwp> If you decide that isn't what you want. Or whatever. And want to return to using Apache. Then just reverse those edits you did to change the port back to port 80 and restart apache.
1617[09:55:52] <rwp> Saying this because I am going to drop offline soon for sleep and I don't want to leave you hanging. :-)
1618[09:58:23] <rwp> SerajewelKS, Looking back on plasmoduck's pastebin replaced-url
1619[09:58:40] <SerajewelKS> rwp: but "file -s" on it said "DOS/MBR boot sector"
1620[09:58:49] <SerajewelKS> rwp: so it's not "just NTFS"
1621[09:58:50] <rwp> Hmm...
1622[09:58:51] <jdhaddon> I saw that, lol. Goodnight (sir/ma'am/other), I look forward to picking your brain
1634[10:02:55] <plasmoduck> so how o I mount that command?
1635[10:02:58] <rwp> Welcome back plasmoduck. I am going to leave you in SerajewelKS' capable hands. Especially since I was failing miserably. Good luck!
1636[10:02:59] <plasmoduck> hey its an exfat artition
1743[11:42:34] <seekr> apollo13: well, I tried that channel last night and couldn't quite get the help I wanted from the person who responded to my questions, for a couple reasons...
1756[11:46:00] <seekr> apollo13: It appears there's been an evolution in the way configuration is done. I did some simple configs back in the NTSC days - a couple decades ago. But Apache splits configuration over a number of files, it seems. I don't know how much is Apache and how much is Debian.
1757[11:46:08] *** Quits: morsicus (~morsicus@replaced-ip) (Remote host closed the connection)
1761[11:46:34] <seekr> apollo13: I'm actually having to do the setup on an Ubuntu server -- but I think it uses the Debian structure.
1762[11:46:50] <apollo13> seekr: so please ask in #ubuntu :)
1763[11:47:10] <seekr> Well, I did, but didn't get any responses.
1764[11:47:15] <apollo13> ubuntu does have slightly different patches, we do not know what ubuntu changed or did not change, so it would be pretty much guesswork
1770[11:49:10] <seekr> apollo13: From one (rather old) printed source, I got the advice to stick a line into a config file: LoadModule rewrite_module modules/mod_rewrite.so
1771[11:51:20] <seekr> apollo13: However, I'm not sure if it's needed if there's a "rewrite.load" symlink in apache2/mods-enabled .
1781[11:53:52] <seekr> okay - I know there's an IRC tradition to ask those seeking advice "What's your real question?," so I'll save you the bother of asking :) I'm transfering a Joomla!-based web site from a shared hosting server to one running on a cloud in a VM. It appears that URL rewriting is not being done. I'm trying to remedy that situation.
1782[11:53:57] <seekr> apollo13: ^^
1783[11:54:15] <apollo13> gotta leave office, sorry. enough for today :)
1936[13:38:56] <uioh> Hello. I just deleted an important file. I wrote a bad rsync command wherein a file on a usb and a file a computer sync to eachother one after the other. I modified the HDD file and plugged in the usb to save the file, but the usb sent it's file to the HDD and overwrote the new stuff.. is it possible to save the overwritten file?
1943[13:44:01] <at0m> uioh: no need to prepend the 'lxterminal -e' if you just want to run the rsync. or throw the line(s) in a bash script, and run that..
1950[13:46:42] <at0m> feel free to keep it, but it's somewhat superfluous unless you want to see the output of the command while it's running. you can also send the output to a file for verification in case something gone wrong
1959[13:51:00] <at0m> any disk write your does now may overwrite the sector(s) containing the deleted info. quickly drop to single user mode (log out from LX, log in as root to tty Ctrl-Alt-F1) and remount the partition read-only. something like mount -o remount -r
1960[13:51:07] <at0m> *your machine does
1961[13:51:44] <at0m> mount -o remount -r /mountpoint
1962[13:52:27] <at0m> hmm or was it mount -o remount,ro /mountpoint
1967[13:55:18] <at0m> but you're on it now, right, so you can proceed in the running OS. just make sure you got somewhere for extundelete to write to, like that usb stick
1968[13:55:54] <uioh> I'm not well versed in linux... could you give me very precise instructions? How to install? How to prevent overwriting?
1969[13:56:11] <uioh> at0m: Right now I'm on the machine just after overwrite.
1970[13:56:25] <uioh> at0m: What commands should I use?
1986[14:00:23] <at0m> uioh: this may kick you off of irc, not sure, might even freeze your machine. so you want to prevent any further writes to the disk. for that, Ctrl-Alt-F1 and log in as root there. then apt install extundelete (ouch there's your write to disk) fsfreeze -f , or mount -o remount,ro /mountpoint (possibly / if you use just one partition)
2001[14:09:03] <at0m> if anyone else cares to hop in on this, feel free ;p
2002[14:10:05] <Kadigan> Hey - what service should I install to have a time service that is Windows-compatible? >_<
2003[14:10:38] <Kadigan> Google is being unhelpful. It's either telling me how to set up time sync on Windows (which I know), or how to set up a time SERVER on Windows (which is not what I want).
2004[14:10:43] <uioh> at0m: is the path to file just /hom/USER/documents...........?
2005[14:10:50] <uioh> */home
2006[14:11:05] <uioh> Or do I have to put something longer?
2009[14:12:21] <blackflow> Kadigan: you can use systemd-timesycnd for ntp and then see set-local-rtc option of timedatectl(1) manpage. I am assuming you're asking about a dual boot situation and Windows keeping hw clock in local time instead of UTC?
2013[14:13:26] <at0m> uioh: ""Note that /path/to/file is the original path of the file in the /dev/sda1 partition!"" <- so the full path of the original file name, including the file name
2014[14:13:46] <uioh> at0m: Is this true for encrypted partitions?
2015[14:14:06] <uioh> at0m: I'm looking in the disque utility and getting very confused.
2021[14:14:47] <Kadigan> blackflow: not a dual-boot situation... well, yes, but no. Not on the same machine. I have a dual-boot on PC-X where Windows is consistently -1h (PC-X dual-boots with Deb9). I have a second PC, PC-VM, which I have Deb9 on too. I want PC-VM to become my network NTP server. I installed and configured `ntp` and port-forwarded 123 UDP because I thought that would work. It doesn't seem to.
2022[14:15:13] <at0m> Kadigan: apt show openntpd
2023[14:15:28] <at0m> that's a time server
2024[14:15:31] <Kadigan> Okay.
2025[14:15:59] <Kadigan> It does mention 'alternate' packages, ntp is one of them. Still, I assume I should remove `ntp` and go with `openntpd` as you suggest?
2026[14:16:14] <at0m> Kadigan: but beware that if PC-VM is a virtual machine, it will just take time from the host
2068[14:33:52] <Kadigan> Well, thanks. It seems my ntp is working, I just can't get through VBox for some reason. I'm seeking assistance w/ XigmaNAS people. Thanks for your time.
2105[14:53:11] <PaddyF> is there a way to let debootstrap be more verbose?
2106[14:53:49] <uioh> Assumer I learn how to unmount the file, would this command be good? extundelete /dev/mapper/MACHINE--vg-root --restore-directory home/USER/<pathtofile>
2121[14:58:25] <uioh> PaddyF: I don't know how to unmount the partition, and I don't understand how things will still run if I unmount the computer I'm on.
2122[14:58:53] <PaddyF> as long as your operating systems runs from that partition you cant unmount it
2123[14:59:14] <uioh> PaddyF: So I can't use extundelete then?
2142[15:07:57] <Xtyu> So, best way I can describe the situation is: Today when testing a live version of windows on USB, I discovered afterward that my laptop no longer recognized my Hard Drive as having an OS anymore. I get a 'no bootable device' error and that's it. So before I make any mistakes and somehow wipe the contents of my HD, I wanted to confirm how to return everything back to how it was.
2143[15:08:02] <tsglove> uioh, sorry to send more stuff you way, yet remember you have two operations: mounting, and "opening" the encrypted container
2144[15:08:33] <tsglove> Xtyu, you had debian installed previously on the laptop?
2149[15:10:18] *** Quits: JohnML (~john1@replaced-ip) (Remote host closed the connection)
2150[15:10:32] <whislock> uioh: If the data has been overwritten at the physical disk level, no.
2151[15:10:39] <tsglove> uioh, well, ... if you know what you're doing, it's easy. Sorry mate... yet it comes with the territory. If the file(s) is critical, I would clone the disk before proceeding.
2152[15:11:07] <uioh> whislock: It was overwritten with rsync
2155[15:11:22] <tsglove> Xtyu, if the USB drive is no longer there plugged in... I don't know man. I doubt the other OS would have messed-up something.
2156[15:11:39] *** Quits: papbrow (~papbrow@replaced-ip) (Quit: My MacBook has gone to sleep. ZZZzzz…)
2157[15:11:43] <uioh> And now I'm not sure what to do.
2158[15:11:45] <tsglove> uioh, I would clone the disk man.
2159[15:11:46] <whislock> uioh: What filesystem are you using on top of the encrypted container?
2160[15:12:12] <uioh> whislock: I used the debian installer to encrypt the whole disk,
2161[15:12:27] <tsglove> probably ext4 then?
2162[15:12:38] <uioh> tsglove: I don't know....
2163[15:13:41] <uioh> I tried following the replaced-url
2164[15:13:45] <whislock> ext4 writes data in place to existing files.
2165[15:13:59] <uioh> whislock: What does that mean?
2166[15:14:01] *** debhelper sets mode: +l 1464
2167[15:14:07] <uioh> whislock: Is it lost?
2168[15:14:09] <tsglove> If the data is critical, clone the disk dude. If it is NOT critical... then proceed.
2169[15:14:20] <whislock> uioh: In all likelihood, yes.
2170[15:14:27] <uioh> whislock: Shit.
2171[15:14:38] <whislock> extundelete is meant to find remnant data from a file that has been deleted, i.e. unlinked.
2172[15:14:54] <whislock> Writing data into an already-existing file writes that data directly to the already-allocated regions of the disk.
2173[15:15:08] <tsglove> In that case ^ ^ ^ data is a gonner.
2213[15:21:02] <whislock> Deleting a file is simple. You're telling the filesystem "I'm not using this space anymore." Overwriting a file means writing new contents.
2214[15:21:14] <whislock> Once you hit that "new contents" part, the old data is gone.
2215[15:21:16] <tsglove> uioh, I hope it wasn't your bitcoin waller
2344[16:20:17] <halt> Hey all, I'm trying to identify outgoing network connections on my debian setup, to be more specific DNS queries run by the system, what would be the best way to capture the process ( executable ) which made the query, I can setup network listener for both TCP and UDP and capture all but how can I find out what process made that call ?
2345[16:21:18] <halt> I'm right now looking at the /proc/net/tcp and the /proc/net/tcp6 files and check that, as far as I understand it should be visible there, but not sure that's the right way to go around this
2350[16:23:54] *** Quits: i1nfusion (~i1nfusion@replaced-ip) (Remote host closed the connection)
2351[16:23:56] <halt> I was considering from both the iptables extra logging scripts, or from the angle of running a custom DNS server but those options look like non native, I mean the OS should be able to give that info a bit more diractly
2376[16:34:46] <halt> the tcp / udp aka network level part is easy I already have a script for that, just need to track back the query to a process / executable
2396[16:41:59] <whislock> Use a local resolver to cache the queries themselves, and something like auditd to log which processes make queries to the resolver.
2414[16:44:53] <whislock> halt: If you're that paranoid, use iptables to rewrite outbound DNS queries to silently direct them toward the local resolver.
2424[16:48:18] *** Quits: abff (~abff@replaced-ip) (Quit: No Ping reply in 180 seconds.)
2425[16:48:18] <halt> Oh yeah I'm paranoid, I'm not shy to deny that, and yeah auditd, might be a good options, I tried to configure a rule for this but given up at some point after few failed attempt, but then I need to revisit that, and check it again
2437[16:57:17] <xparanoik> curious to know if there's such a feature in apt to change a "default" package, let's say apache2 to nginx. any suggestions on what to search for?
2460[17:12:04] <uio> Hi, I'm back still steaming about a lost overwritten file. Does if I have file important.txt on a usb and on my hdd, and I modify it on my HDD, will rsync -avh /dev/usb/important.txt /home/user/Documents/important.txt overwrite or rather delete the file on my HDD?
2490[17:22:28] <PaddyF> the software cant delete a file in such a way that you can recover it cause it does not know when your actions are final and when not etc.
2517[17:29:32] <fearnothing> hi, I would like some help with an RDP problem. I recently updated jessie->stretch and after doing so, RDPing from my Windows system doesn't work correctly
2525[17:31:44] <Fah> uio: depending on what you're doing, put the disk in ro mode. You might be able to recover some raw blocks off the device with debugfs, but in general its hard going and you're goig to really need to understand a lot about the underlying file system. The reason there's no simple how to is because there's no simple solution.
2539[17:41:00] <uio1> Away from overwritten files to more interesting things : ThinkPad fans. My fan1 is usually running at around 3K Tr/mn. and I can hear it, start up and slow down. Is this normal?
2544[17:45:22] <jhutchins> uio1: It depends. What activities cause it to speed up? There are fan control programs, although it's often controlled by the BIOS. It should run hard when you're doing heavy computation or graphics.
2551[17:49:38] <uio1> jhutchins: Yes, its from 2008. But there is a pulsing noise... is there an audio file paster?
2552[17:49:47] <jhutchins> uio1: You should probably read some online documentation on rsync and how it works. Yes, if you tell it to copy a shorter file over a longer one you will loose the longer one, and are unlikely to recover the missing data.
2553[17:50:03] <jhutchins> uio1: avh will not delete files.
2558[17:52:27] <jhutchins> uio1: I am not troubleshooting your fan. Follow Fah's advice, install lm-sensors and some GUI monitoring display and figure it out for your self. Clean the system.
2703[19:11:06] <nedstark> the machine learning community is not a big one
2704[19:11:14] <whislock> !offtopic
2705[19:11:14] <dpkg> #debian is primarily a support channel for Debian users. Please keep the discussions in #debian on-topic and take longer discussions and non-support questions to #debian-offtopic. Imagine the chaos if each of the hundreds of people in the channel felt the need to wander off topic for a few minutes every day.
2706[19:11:17] <nedstark> you're not even in the channel
2906[21:14:22] <znoteer> I frequently copy and paste between xterms by selecting text with the left mouse button, then moving the mouse cursor to where I want to paste, and pasting with a middle button click. When I'm using vim, however, it doesn't seem to work. Instead of pasting what I've selected in another window, it pastes something else. The opposite is true, when I select in vim then try to paste in another window, it
2907[21:14:24] <znoteer> doesn't paste what I just selected. Does anyone know how to make copying and pasting with mouse work with vim?
2909[21:15:13] <greycat> vim in stretch broke with tradition. In order to get back to normal, you must create a ~/.vimrc file to prevent the system-wide equivalent from being used.
2910[21:15:53] <greycat> Either that, or try to adapt to the new system-wide default vimrc file. I think you can hold Shift to use the mouse when that one is in effect, but I'm not 100% sure.
2937[21:30:06] *** aloo_shu is now known as MarxXXL
2938[21:31:42] <SerajewelKS> greycat: you can hold shift in most terminals. i've used some that use alt though. so it depends on the terminal being used.
2939[21:32:16] <greycat> I've had a .vimrc for a long time, so it didn't affect me at first, until I tried to vi something as root.
2940[21:32:37] <gumnos> Ahoy. Found an old Etch machine (decent specs) in a closet and would like to bring it up to Stable. However apt is giving me KEYEXPIRED errors and tips for updating the keyring (`apt-key adv --keyserver keys.gnupg.net --recv-keys ${KEY}`) doesn't seem to resolve
2941[21:33:09] <SerajewelKS> the "shift" trick is a terminal thing, not a vim thing. it tells the terminal not to pass the mouse interaction information to the program in the terminal.
2942[21:33:11] <gumnos> What do I need to do to get apt to accept the keys?
2943[21:33:31] <SerajewelKS> so it's up to the terminal which modifier to use for that, if it even supports that
2960[21:37:46] <gumnos> Reading the man-pages on apt and apt-get, I'm not seeing anything referencing how to ignore expired keys. Is there some undocumented option I could be using to ease this process?
2979[21:43:08] <joepublic> the oldest system I have here is wheezy
2980[21:43:36] <SerajewelKS> i think apt prompts you anyway whether you want to install the unauthenticated packages and you can just say yes
2981[21:43:42] <SerajewelKS> IIRC the flag just skips that prompt
2982[21:45:05] <gumnos> Okay, I'll see what I can accomplish to bring this box to a more modern rendition. It was an SFTP/FTP file drop with some SMB shares, and I'm hoping to figure out any other service that it's running in there too.
2984[21:45:52] <SerajewelKS> please do the backup first. you will probably run into some problems and you may need to restart the upgrade if you get too stuck.
3154[23:27:17] <znoteer> jhutchins: I just tried vim.nox. It behaves the same as vim.basic. Copy/paste with the mouse doesn't work unless you hold down Shift, now.