| golinux | <systemdlete> Or is doing an update enough to get 2.1? | 00:02 |
|---|---|---|
| golinux | Yes. update and upgrade like you would normally. You don't have to do anything special. | 00:02 |
| systemdlete | so no dist-upgrade needed -- that would be to go from, say, ascii to beowulf then, right? | 00:03 |
| tme520 | Hi there | 00:05 |
| golinux | I'm not sure actually. I think it would depend what's in your sources.list. Can't you check to see what release you're running? I'd use hardinfo for that/ | 00:05 |
| systemdlete | This is interesting. My ascii-vm now boots to a blank screen; logging out of the last session left me with a blank/hung screen. This is in a vbox VM. Didn't have that problem previously. | 00:07 |
| systemdlete | the dm is slim, btw | 00:07 |
| systemdlete | I have not changed slim options, or any other desktop config recently in that VM. I did, however, try to force re-install slim to see if that might make the blank boot up go away and bring back the dm | 00:08 |
| * systemdlete scratches his head, trying to recall if he made any recent mods but nothing comes to mind. | 00:08 | |
| systemdlete | I also note that postfix did not come up on the host, also running ascii, when I rebooted the host system 3 days ago. | 00:08 |
| systemdlete | (still looking at that) | 00:09 |
| systemdlete | running startx from command line (in tty2) works fine. I get my normal xfce desktop. | 00:10 |
| systemdlete | ... switching back to tty7, that screen is still blank and hanging | 00:11 |
| systemdlete | So something having to do with slim I guess | 00:11 |
| systemdlete | Looks like some issue with AIGLX? drmDropMaster failure | 00:13 |
| golinux | Look at the 2.0.0 release notes. They must be around somewhere . . . | 00:14 |
| golinux | I don;t have time to dig atm | 00:14 |
| golinux | Will do so later. | 00:14 |
| systemdlete | I'm up-to-date, which I thought you said brings me to 2.1 ascii? | 00:14 |
| systemdlete | OK, thanks. | 00:15 |
| golinux | It should | 00:15 |
| systemdlete | so 2.0.0 release notes would be somewhat non-relevant now? | 00:15 |
| systemdlete | I mean, the VM has worked for a long time... ever since I built that vbox | 00:16 |
| golinux | I'm really the wrong person to tackle an issue like this. I'm more of the loibrarian. | 00:16 |
| golinux | -o | 00:16 |
| systemdlete | np, thanks for your input | 00:16 |
| systemdlete | googling this issue; seems others have encoutnered it also | 00:17 |
| golinux | The backend stuff that would affect slim I think are also in the 2.1 release notes. | 00:17 |
| golinux | I get confused between what's in the release notes and the announcement. | 00:18 |
| golinux | Here you go. I'll bbl http://files.devuan.org/devuan_ascii/Release_notes.txt | 00:19 |
| golinux | Look for: - Session management and policykit backends | 00:19 |
| systemdlete | Yeah, I found that already. But why is it saying that "What's new" includes the 4.9.0-9 linux-image? I only have -7 and -11 images | 00:20 |
| systemdlete | nvm. I read that too quickly. | 00:20 |
| systemdlete | I have -9 and -11, indeed. | 00:21 |
| systemdlete | aha... I see what you are saying. The Release Notes do include stuff for 2.1, and they do say that Kernel Mode Setting is required... I'll look at that | 00:25 |
| systemdlete | Now it is working again. Needs_root_rights=yes added to the Xwrapper.conf file, as per release notes. Thank you for pointing this out. | 00:33 |
| systemdlete | I usually ignore the release notes for most distros because, historically, I have rarely had issues. | 00:33 |
| systemdlete | After reboot, however, the problem re-appears. No DM (i.e., no slim). the wrapper file still has the change I made. | 00:36 |
| systemdlete | ??? | 00:36 |
| systemdlete | I was wrong. It DOES come up... after a long recess! | 00:49 |
| systemdlete | it does NOT, however, save my resolution. In fact, when I switch from full screen to window mode, it goes back to 800x600. When I boot, I get 800x600 also. | 00:51 |
| systemdlete | nomodeset does not help really | 00:53 |
| systemdlete | It still takes eternity for the DM to appear | 00:53 |
| systemdlete | well, nomodeset DOES restore my resolution in windowed mode. But it still takes forever to come up. | 00:56 |
| systemdlete | dmesg shows that there's a drm error about 60 seconds after the previous messages, and that's about how long the delay is | 01:10 |
| systemdlete | nvm my crack about drm error: That was on my host devuan. | 01:31 |
| systemdlete | In my VM, when I logout of xfce (the polite way), the screen goes to blank for a very long time (indeterminate) | 01:32 |
| systemdlete | If I open a window on, say, tty1 and restart slim, the running slim won't die -- stop fails. | 01:32 |
| systemdlete | but slim does restart successfully, and the DM appears in tty7. | 01:32 |
| tuxd3v | #join #devuan-arm | 01:39 |
| Hurgotron | I upgraded a rather small install of ascii to beowulf, now ssh logins have a delay of 25 seconds. | 01:44 |
| Hurgotron | auth.log says "Failed to activate service 'org.freedesktop.login1': timed out (service_start_timeout=25000ms)" | 01:44 |
| Hurgotron | The web suggests that restarting systemd-logind helps... but I'm on Devuan. :) | 01:45 |
| Hurgotron | anyone? | 01:45 |
| golinux | Doesn't elogind take care that? | 01:47 |
| golinux | I am no expert but check if you have any of that installed | 01:47 |
| Hurgotron | yes it's installed... should it be running? | 01:48 |
| golinux | Read the release notes carefully: http://files.devuan.org/devuan_ascii/Release_notes.txt | 01:52 |
| golinux | The answer is probably in there. | 01:52 |
| golinux | It definitely isn't in my head. :D | 01:53 |
| Hurgotron | well it worked with ascii, without elogind :) | 01:53 |
| Hurgotron | But anyway, I'm on the right track now. | 01:53 |
| Hurgotron | fucking cgroup shit isn't available in my vserver. | 01:54 |
| tuxd3v | hello, does any body has devuan running in the HP Mycloud Home NAS? | 01:55 |
| gnarface | tuxd3v: i don't have one, but what is the issue you're having with it? | 01:59 |
| tuxd3v | gnarface, mine doesn't run devuan :( | 02:00 |
| tuxd3v | I was thinking in.. see if possible to have devuan running there :) | 02:00 |
| tuxd3v | I believe a propeietary module or something was needed, in the pastm don't know if linux kernel already has all what is needed.. | 02:01 |
| gnarface | tuxd3v: oh, i thought you were just having trouble locating the Samba mount point. i don't know about actually installing linux onto one. | 02:01 |
| tuxd3v | but even so, I also need support from the bootloader.. I am not yet getting the dust out of it, but I will when I have devuan running like it should in my RPi1 B v1.0 :) | 02:02 |
| systemdlete | So I killed the slim from tty1. Started it manually, with -d, but now the whole VM seems to be hung. | 02:03 |
| tuxd3v | right now is compiling a kernel... 2 days now, and counting :) | 02:03 |
| systemdlete | Question: Was there any changes in 2.1 having to do with entropy? Some of the entries re slim delays seem to revolve around entropy issues. | 02:03 |
| systemdlete | (entries = google hits) | 02:03 |
| tuxd3v | systemdlete, does you have rng-tools installed ? | 02:04 |
| tuxd3v | if so point it to '/dev/urandom' just to ckack the solution :) | 02:05 |
| systemdlete | no, I tried that. It's a VM, so there's no rng device. But I did install haveged, which was recommended. | 02:05 |
| systemdlete | no help | 02:05 |
| systemdlete | same problem persists | 02:05 |
| tuxd3v | I tried haveged but got rng-tools back, with '/dev/urandom' in /etc/defaults/rng-tools | 02:06 |
| systemdlete | hmmm. In a VM? | 02:06 |
| tuxd3v | for know in kernel 5.3 no support for TRNG on Allwinner H6( Orange Pi one plus.. ), so its the only option.. | 02:07 |
| systemdlete | At any rate, running slim from command line, as root, results in a hang (in a VM at least) | 02:07 |
| systemdlete | (I'll try that, thanks) | 02:07 |
| tuxd3v | but are you sure it has to do with rng? | 02:07 |
| systemdlete | what has to do with rng? The long delay or the slim hang? | 02:08 |
| systemdlete | I googled for hits on long delay on slim and some of the hits suggest entropy issues. | 02:08 |
| tuxd3v | ho.. | 02:08 |
| systemdlete | google "slim: waiting for x server to begin accepting connections" | 02:08 |
| systemdlete | (that's the last message I see in slim.log) | 02:09 |
| gnarface | i think you should be able to get urandom in a VM... it's not a hardware device... | 02:12 |
| tuxd3v | I saw some hist about problems related with glibc 2.38, but we are in 2.24-11 I believe ?! | 02:13 |
| tuxd3v | yeah I sugested puting it in /etc/default/rng-tools | 02:13 |
| systemdlete | gnarface, tuxd3v: I'm trying it now! | 02:13 |
| tuxd3v | hist -> hits | 02:14 |
| systemdlete | tuxd3v: DId you try googling what I did, see above | 02:14 |
| tuxd3v | systemdlete, let us know if that solved the problem | 02:14 |
| systemdlete | oh, I will, believe me... :D | 02:14 |
| systemdlete | noop | 02:14 |
| systemdlete | still hangs at boot | 02:14 |
| tuxd3v | yes, I saw one of the first hits about glibc 2.38 on gentoo | 02:15 |
| systemdlete | I use the term "hang" loosely, of course. Eventually, I do get the login screen | 02:15 |
| systemdlete | look at subsequent hits also | 02:15 |
| systemdlete | some of the suggestions I've seen are to wiggle the mouse a bit during boot... that made me laugh | 02:17 |
| systemdlete | reminds me of that Dilbert cartoon where his boss can't get his "laptop" to work... | 02:18 |
| systemdlete | he tells the pointy-haired boss to "turn it over and shake it" | 02:18 |
| tuxd3v | well, that is in fact related to low entropy | 02:18 |
| tuxd3v | or slow entropy device.. | 02:19 |
| systemdlete | Funny, I didn't have slow entropy until now | 02:19 |
| tuxd3v | because entropy is got by several ways | 02:19 |
| tuxd3v | I don't have problems with my slim | 02:19 |
| systemdlete | are you in a vm? | 02:20 |
| tuxd3v | I use slim and slimlock...I am just amazed by its simplicity, beauty, and functionality | 02:20 |
| tuxd3v | no | 02:20 |
| fsmithred | systemdlete, did you install the full desktop or just the parts you want? | 02:20 |
| systemdlete | ah | 02:20 |
| tuxd3v | I am in a real server :D | 02:20 |
| systemdlete | :p real server... | 02:20 |
| tuxd3v | yeah | 02:20 |
| tuxd3v | if you want I can share my config with you.. | 02:21 |
| tuxd3v | another thing.. | 02:21 |
| tuxd3v | does you have a .Xinitrc file | 02:21 |
| systemdlete | fsmithred: Depends what we are talking about here. This VM was ascii 2.0 and I have updated/upgraded to newest packages which I have been led to believe makes my VM 2.1 ascii now | 02:21 |
| systemdlete | was working under 2.0 --> now not working under 2.1 with (mostly) same software | 02:22 |
| fsmithred | is this the first time you've upgraded since you installed the system? | 02:22 |
| systemdlete | *other software | 02:22 |
| golinux | <systemdlete> Question: Was there any changes in 2.1 having to do with entropy? | 02:22 |
| systemdlete | no, not by any means. I have been updating pretty often | 02:22 |
| fsmithred | then you were already there | 02:22 |
| systemdlete | ok | 02:22 |
| golinux | There is in beoowulf iirc | 02:22 |
| fsmithred | the only reason to call it 2.1 is because there are updated isos | 02:22 |
| systemdlete | right | 02:23 |
| systemdlete | I even tried to force re-install slim. Which it did, but to no avail. | 02:23 |
| fsmithred | yeah, in beowulf you can get delays with openssh-server | 02:23 |
| fsmithred | make sure you've got all the right libpolkit packages | 02:24 |
| tuxd3v | hehe, my RPi1 is compiling a kernel for 2 days , and I suspect a 3rd one is comming.. | 02:24 |
| systemdlete | I also think that those hits are talking about much newer kernels, like 5.x | 02:24 |
| golinux | systemdlete: Do you have a mixed system? | 02:24 |
| tuxd3v | My "desktop", compiled a kernel in the moments I were here speaking with systemdlete :D | 02:24 |
| systemdlete | tuxd3v: impressive | 02:24 |
| systemdlete | golinux: ? | 02:25 |
| fsmithred | is there any beowulf in your sources.list? | 02:25 |
| systemdlete | shouldnt be | 02:25 |
| tuxd3v | the thing is.. I doesn't wanted the kernel compiled, just wanted the process to download it, but it downloaded and compiled it without I notice it.. | 02:25 |
| systemdlete | The only VM I've done that in is my beowulf experimental VM. | 02:25 |
| tuxd3v | this machines are bacoming fast... | 02:26 |
| tuxd3v | bacoming- > becoming | 02:26 |
| systemdlete | tuxd3v can have his pie and eat it too | 02:26 |
| systemdlete | s/pie/pi/ | 02:26 |
| systemdlete | what is a "mixed" system? | 02:27 |
| fsmithred | mixed stable/testing or other | 02:27 |
| golinux | ascii/beowulf | 02:27 |
| systemdlete | mixed-up, probably. But that has more to do with the operator I am afraid... | 02:27 |
| systemdlete | ah. | 02:27 |
| systemdlete | No | 02:27 |
| fsmithred | really, check the polkit shit | 02:27 |
| golinux | I second that | 02:27 |
| systemdlete | As I said, my only references to beowulf are in an ascii-cum-beowulf VM I have been experimenting with occasionally. That's a different VM though. | 02:28 |
| systemdlete | Oh what fun. Check the polkit shit... | 02:28 |
| * systemdlete goes off, not to be seen for years and years | 02:28 | |
| fsmithred | dpkg -l | egrep "consolekit|elogind|policykit|polkit|libpam" | 02:28 |
| fsmithred | and past result | 02:28 |
| fsmithred | paste | 02:28 |
| golinux | You have ascii in your sources.list or testing? | 02:28 |
| systemdlete | I can tell you that there is no elogind here, only consolekit, for compat with xfce, as per the instructions in release notes | 02:29 |
| golinux | oops or dtable | 02:29 |
| systemdlete | hold on | 02:29 |
| golinux | stable | 02:29 |
| golinux | If you have stable your are pulling debian pkgs from Buster. | 02:30 |
| systemdlete | dpkg -l | egrep "consolekit|elogind|policykit|polkit|libpam" ---> http://paste.debian.net/1117962/ | 02:30 |
| systemdlete | sources.list --> http://paste.debian.net/1117963/ | 02:31 |
| fsmithred | install libpam-ck-connector | 02:31 |
| fsmithred | pretty sure | 02:31 |
| systemdlete | /etc/apt/sources.list.d/devuan.list is all commented out | 02:32 |
| systemdlete | oh shit | 02:32 |
| systemdlete | never mind those | 02:32 |
| * systemdlete has finger-checked again | 02:32 | |
| systemdlete | right pew, wrong church | 02:33 |
| systemdlete | those are for my host devuan | 02:33 |
| systemdlete | (sorry!) | 02:33 |
| systemdlete | dpkg -l |grep etc ---> http://paste.debian.net/1117965/ | 02:36 |
| systemdlete | http://paste.debian.net/1117966/ | 02:36 |
| systemdlete | (sources.list) | 02:36 |
| fsmithred | they're the same | 02:37 |
| fsmithred | and you don't have the problem in the host system? | 02:37 |
| systemdlete | ehhh. No. Though I've had some OTHER interesting problems lately. | 02:37 |
| fsmithred | one more thing to check. It's unlikely, but maybe some backports packages slipped in. | 02:38 |
| systemdlete | When this ascii VM locked up a day or two ago, due to some problem with a web page, it caused the hostkey to stop working, even though I could ssh in from my testbox to the host | 02:38 |
| fsmithred | dpkg -l |grep bpo | 02:38 |
| fsmithred | will give some false positives. Look for ~bpo in the version | 02:39 |
| systemdlete | http://paste.debian.net/1117967/ | 02:39 |
| systemdlete | yeah, lots | 02:40 |
| systemdlete | remmina libpolkit | 02:40 |
| fsmithred | did you intend to install all those from backports? | 02:40 |
| systemdlete | not sure. Remmina, maybe. Long time ago. | 02:41 |
| systemdlete | is dovecot coming from backports also? | 02:41 |
| fsmithred | yes | 02:41 |
| systemdlete | oh | 02:41 |
| systemdlete | is there a normal one in the other repos? | 02:42 |
| fsmithred | all the ones with ~bpo in the version | 02:42 |
| systemdlete | if not, I don't mind using package called "imap" instead | 02:42 |
| golinux | BTW us.deb.devuan.org is not recommended as there is no us mirror | 02:42 |
| fsmithred | no remmina in ascii | 02:43 |
| systemdlete | ah, well, I can probably remove remmina | 02:43 |
| systemdlete | fsmithred: Do you suspect an incompatibility causing this slim issue? | 02:44 |
| systemdlete | Or is your concern elsewhere? | 02:44 |
| fsmithred | it's possible | 02:44 |
| systemdlete | ok | 02:44 |
| fsmithred | I don't know what all those backported polkit packages do | 02:45 |
| golinux | I was thinking that | 02:45 |
| * systemdlete makes remmina go *POOF!* | 02:45 | |
| fsmithred | oh | 02:45 |
| fsmithred | policykit-1 in ascii-security is the bpo version | 02:45 |
| fsmithred | try installing libpam-ck-connector | 02:46 |
| systemdlete | eh... I'm not getting something here. dpkg -l is still showing dovecot-core and dovecot-imapd but I can't remove them because apt-get thinks they're not installed? | 02:48 |
| systemdlete | I was able to remove remmina no problems | 02:49 |
| Xenguy | Hey folks, I saw an email about a point release for Ascii, but no updates have shown up yet? Any ideas why? | 02:49 |
| Xenguy | My mirror is deb.devuan.org | 02:50 |
| Xenguy | (as recommended, I believe) | 02:50 |
| Xenguy | systemdlete: dpkg -r ? | 02:51 |
| gnarface | i don't know anything specific about the point release, but sometimes mirror propagation takes a few hours | 02:51 |
| gnarface | how long ago was it announced? | 02:51 |
| Xenguy | gnarface: Yeah, probably just need to be patient | 02:51 |
| fsmithred | Xenguy, the new isos have been up for a couple days | 02:51 |
| * Xenguy checks the email timestamp... | 02:51 | |
| fsmithred | if you've upgraded recently, you're already there | 02:51 |
| golinux | Just do a normal upgrade and you should be there. Those pkgs have been arounf for a while so you | 02:52 |
| golinux | should have them | 02:52 |
| systemdlete | nvm. Those are just the config files. I want to save those. | 02:52 |
| golinux | That mind meld thing again. | 02:52 |
| Xenguy | Hrm, so maybe this is just a false alarm | 02:52 |
| golinux | There is no alarm | 02:52 |
| Xenguy | OK thanks, I'll chill | 02:52 |
| tuxd3v | sustemdlete, have you tested what we spoke above? | 02:53 |
| golinux | There is if you are installing for the first time. The apt security bug fix is in the new isos etc, | 02:53 |
| se7en | I just read the email stating that there is an ASCII upgrade, v2.1 | 02:53 |
| se7en | Will `apt full-upgrade` upgrade my system? | 02:54 |
| Xenguy | Meanwhile, I miss y'all (you too furrywolf) but I've been kidnapped by HR (the thieving bastards!)... | 02:54 |
| golinux | Probably just a normal upgrade would do. | 02:54 |
| Xenguy | It'll all be over by Wednesday, and a bloody Amen to that | 02:54 |
| se7en | I usually run `apt update && apt full-upgrade` to ensure I don't have broken packages anyway | 02:55 |
| se7en | Also, any plans to host Libressl on apt? | 02:55 |
| se7en | I'm a bit tired of compiling it, and then in apt openssl often installs and replaces it | 02:55 |
| fsmithred | make a deb package so it won't do that | 02:56 |
| se7en | Well, I've never made a .deb before so I'm a bit afraid that I'll do it wrong | 02:57 |
| gnarface | se7en: just use checkinstall. it is easy | 02:57 |
| se7en | Will look into, gnarface | 02:57 |
| Xenguy | Thanks gnarface fsmithred golinux , I think I'm good | 02:58 |
| Xenguy | ^^ Checkinstall is awesome | 02:58 |
| gnarface | Xenguy: if you need to be sure, i'd say just check the last few things in your /var/log/apt/history.log against the versions showing up in pkginfo.devuan.org currently | 02:59 |
| gnarface | but yea if you recently ran an "apt-get upgrade" and got packages, you probably have all the updates already | 02:59 |
| Xenguy | Right on (just after I kick HR's ass ; -) | 02:59 |
| Xenguy | gnarface: I think I just misinterpreted what a "point release" is, so I'm thinking it's all good now | 03:00 |
| systemdlete | sorry, tuxd3v, I forgot what that was... | 03:00 |
| * Xenguy goes up in a puff of smoke... | 03:01 | |
| systemdlete | I remember you wanted me to test something, but I think it scrolled off | 03:01 |
| systemdlete | fsmithred: Still working on removing those policykit stuff from backports | 03:13 |
| systemdlete | problem is, it wants to remove a good chunk of my system with those packages | 03:14 |
| systemdlete | Is there a way to safely downgrade? | 03:14 |
| systemdlete | to pick up those packages from the regular repos | 03:14 |
| gnarface | not really, at this point. downgrades aren't tested upstream. but if you purge all the backport packages it will probably be ok | 03:15 |
| gnarface | if you sucked in the wrong thing though, like say, a newer version of apt that fundamentally restructured your dependency tree.... then there may be no going back | 03:16 |
| systemdlete | "you can checkout any time you like, but you can never leave..." | 03:17 |
| gnarface | yea, something like that. keep backups. | 03:17 |
| systemdlete | oh yes. | 03:17 |
| systemdlete | I do. | 03:17 |
| systemdlete | I backup my backups to the cloud. | 03:17 |
| systemdlete | I'll have to try it at the command line because it wants to remove dbus | 03:18 |
| systemdlete | I'm afraid that'll choke any desktop I'm running like xfce | 03:18 |
| gnarface | it's reasonable to expect it may remove some things you will want to re-add after. just be careful that if it removes the kernel you don't reboot before installing another :) | 03:18 |
| systemdlete | I don't think the kernel was one, no | 03:19 |
| systemdlete | but udisks2 is one it wants to remove | 03:21 |
| fsmithred | systemdlete, you should not remove them. They are from ascii-security | 03:21 |
| systemdlete | but they are ~bpo | 03:22 |
| fsmithred | apt-cache policy policykit-1 | 03:22 |
| fsmithred | don't paste the result. I already know it. | 03:22 |
| fsmithred | Did you try adding libpam-ck-connector? | 03:22 |
| gnarface | well, you don't actually need udisks2 to boot | 03:23 |
| gnarface | it's for gui drive mounting tools | 03:23 |
| systemdlete | well, I'd replace it as soon as I have the rest removed | 03:23 |
| systemdlete | ok, connector is installed. Reboot? | 03:24 |
| fsmithred | they're gonna come back on the next upgrade | 03:24 |
| fsmithred | even if you comment out the backports repo | 03:24 |
| systemdlete | (I have commented out the backports already, and updated) | 03:24 |
| fsmithred | reboot or 'init 1' | 03:24 |
| systemdlete | slim has difficulty shutting down also. Apparently it cannot stop X | 03:25 |
| fsmithred | I'm not surprised. | 03:25 |
| systemdlete | wow. | 03:26 |
| systemdlete | fixed it, fsmithred! fixed it! | 03:26 |
| fsmithred | yup | 03:26 |
| systemdlete | slim came right up | 03:26 |
| * systemdlete tries it again, just for fun | 03:26 | |
| systemdlete | uh-oh | 03:27 |
| systemdlete | spoke too soon. | 03:27 |
| systemdlete | after reboot, back to same fun-and-games... :( | 03:27 |
| systemdlete | crap. | 03:28 |
| * systemdlete wonders if anyone else in this channel is having as much "fun" as he is with an ascii 2.1 upgrade in a VM | 03:29 | |
| fsmithred | nope | 03:29 |
| systemdlete | VM? | 03:29 |
| fsmithred | yeah, most of my test systems are in vm | 03:29 |
| systemdlete | so... | 03:30 |
| gnarface | so what did you do!? | 03:30 |
| systemdlete | As the song suggests, I don't "mess around with (sl)im" | 03:30 |
| systemdlete | I guess I broke it. | 03:30 |
| gnarface | hehe well that part is reasonably obvious. the real question is how, specifically? | 03:31 |
| gnarface | the trick is being able to retrace your steps | 03:31 |
| systemdlete | It all started... | 03:31 |
| systemdlete | I had a bad childhood... | 03:31 |
| gnarface | woah there | 03:31 |
| gnarface | not that far back | 03:31 |
| fsmithred | you only need to go back as far as root's history | 03:32 |
| systemdlete | :D | 03:32 |
| gnarface | probably just back to the last thing you installed/upgraded, or maybe the thing right before that | 03:32 |
| * systemdlete thinks | 03:32 | |
| gnarface | something from backports, or a custom kernel maybe? | 03:32 |
| systemdlete | no custom kernels | 03:32 |
| fsmithred | dovecot? | 03:32 |
| systemdlete | backports, the ones you know about. | 03:32 |
| systemdlete | Ah! | 03:32 |
| gnarface | any custom builds, 3rd party packages or weird non-stock configuration edits should be suspect... | 03:32 |
| systemdlete | now dovecot I've had some problems with | 03:32 |
| systemdlete | which makes me wonder... | 03:33 |
| * systemdlete disables dovecot and reboots... | 03:33 | |
| fsmithred | isn't that some fedora thingy? | 03:33 |
| systemdlete | dovecot? | 03:34 |
| fsmithred | nm, I looked it up | 03:34 |
| systemdlete | idk really | 03:34 |
| gnarface | it's in debian but it's kinda messy | 03:34 |
| systemdlete | it was the first imap server I ever worked with | 03:34 |
| fsmithred | I was thinking of dracut | 03:34 |
| systemdlete | oh, no. not dracut | 03:35 |
| gnarface | it's kinda general purpose imap thing but if you start actually trying to generally repurpose it, things can get weird | 03:35 |
| systemdlete | In fact, I rarely touch tools like those | 03:35 |
| gnarface | oh hmm.... imaps starving entropy was a thing i had an issue with i think once... | 03:36 |
| systemdlete | I can tell you this (maybe this would be helpful?): The last message I see in slim.log is that it is waiting on X to start receiving connections | 03:36 |
| systemdlete | well disabling dovecot helped zero | 03:36 |
| gnarface | yea i saw you mention that earlier. i silently concurred with someone else's hypothesis that it could be a entropy starvation issue | 03:36 |
| systemdlete | how cruel! | 03:37 |
| fsmithred | install haveged? | 03:37 |
| fsmithred | I had to do that for live isos that have openssh-server | 03:37 |
| gnarface | well, switching to /dev/urandom should have worked too | 03:37 |
| gnarface | not securely, but it would have worked | 03:37 |
| fsmithred | what do you switch? | 03:37 |
| systemdlete | yeah, right now, though, I have rng-tools because tuxd3v wnated me to try that, pointing to /dev/urandom, etc | 03:38 |
| systemdlete | but I can go back to haveged | 03:38 |
| gnarface | fsmithred: anything that would otherwise read from /dev/random as a source | 03:38 |
| gnarface | fsmithred: it's a ugly quick&dirty hack but you basically can just symlink /dev/random to /dev/urandom and you should no longer get any blocking while waiting for entropy | 03:39 |
| systemdlete | gnarface: removed rng-tools | 03:39 |
| fsmithred | thanks | 03:39 |
| fsmithred | for the live isos I had to make sure haveged started before ssh | 03:39 |
| systemdlete | what is the entropy being used for anyway? I mean, why does slim demand it? | 03:40 |
| fsmithred | no clue | 03:40 |
| gnarface | yea that's a good question | 03:40 |
| systemdlete | ok, installing haveged | 03:40 |
| * systemdlete usually asks pretty good questions. It's just that the answers to them are kind of, well, you know. | 03:40 | |
| gnarface | does slim support network authentication or something? | 03:41 |
| gnarface | could it be prepping an encrypted socket for something? | 03:41 |
| gnarface | i don't even know, but there are very few rational excuses | 03:42 |
| gnarface | it could be hanging on a network timeout of some sort too though, unrelated to encryption handshakes... | 03:43 |
| systemdlete | ok rebooted with haveged installed and enabled. Still slow to get to slim | 03:43 |
| gnarface | have you timed it yet? | 03:44 |
| systemdlete | I'm wondering if maybe I should just rebuild the system from a 2.1 ISO. I could restore my home and system files. | 03:44 |
| systemdlete | no. But I think it takes about a minute | 03:44 |
| systemdlete | I can time it. | 03:44 |
| fsmithred | a minute? | 03:44 |
| systemdlete | from the moment it goes blank | 03:44 |
| gnarface | if it takes the same amount of time every time within less than a second, that is evidence | 03:44 |
| fsmithred | oh that. | 03:44 |
| fsmithred | I thought you meant the install takes a minute. | 03:45 |
| gnarface | if it's exactly 60 seconds every time, then it's *almost certainly* a network timeout issue | 03:45 |
| gnarface | (possibly related to DHCP and your MTA) | 03:45 |
| systemdlete | i disabled dovecot | 03:45 |
| gnarface | dovecot isn't the MTA | 03:45 |
| systemdlete | not using dhcp or dhcpcd | 03:45 |
| fsmithred | what's in /etc/network/interfaces? | 03:45 |
| systemdlete | postfix | 03:45 |
| fsmithred | auto or allow-hotplug? | 03:45 |
| systemdlete | (doing timing test) | 03:46 |
| gnarface | systemdlete: did you disable postfix too as a test? make sure no alternatives are installed either, like exim4 or sendmail... | 03:46 |
| systemdlete | 02:10 | 03:49 |
| systemdlete | not yet, doing now, then timing again | 03:49 |
| gnarface | hmm. 2:10 is long for a network configuration issue based timeout | 03:49 |
| systemdlete | well, some of that might be something else, although 10 seconds for that other thing is kinda long also | 03:50 |
| systemdlete | let me get you another timing... | 03:50 |
| systemdlete | 2:10 again | 03:53 |
| systemdlete | and both postfix and dovecot were disabled this time | 03:54 |
| systemdlete | what if I try a differnt DM? | 03:54 |
| systemdlete | just to see if it might be slim-related? | 03:54 |
| gnarface | it's a good idea | 03:54 |
| gnarface | you could try without a DM entirely too | 03:54 |
| systemdlete | well, I've done that. | 03:54 |
| systemdlete | the login comes right up | 03:54 |
| tuxd3v | humm | 03:55 |
| systemdlete | sorry, no | 03:55 |
| tuxd3v | gnarface is right, it could | 03:55 |
| systemdlete | I mean, the desktop comes right up | 03:55 |
| tuxd3v | can you share your '/etc/slim.conf' | 03:57 |
| tuxd3v | ? | 03:57 |
| systemdlete | I don't care if it's not compat with my desktop b/c I'm not actually going to try to login, right? | 03:57 |
| systemdlete | http://paste.debian.net/1117975 | 03:57 |
| systemdlete | I dont' recall changing it, but I might have | 03:58 |
| systemdlete | s/dont'/don't/ | 03:58 |
| systemdlete | gnarface: which DM should we try? | 03:58 |
| * tuxd3v ... I made a builder so that I could be in charge, clicking the buttons..now my builder builds kernels without me doing nothing...boring | 03:58 | |
| systemdlete | (tuxd3v: Typical Marxian phenomenon, where the worker becomes alienated from his work) | 03:59 |
| systemdlete | sddm? | 04:00 |
| systemdlete | xdm? | 04:00 |
| systemdlete | lightdm? | 04:00 |
| fsmithred | lightdm should work | 04:00 |
| systemdlete | ok | 04:00 |
| fsmithred | I need to sleep. | 04:01 |
| fsmithred | good luck with it. | 04:01 |
| systemdlete | thank you so much again, fsmithred. | 04:03 |
| tuxd3v | fsmithred, good night | 04:03 |
| systemdlete | g'nite, fsmithred | 04:04 |
| tuxd3v | systemdlete, take my config: https://paste2.org/C8PmFkyx | 04:04 |
| tuxd3v | substitute tuxd3v by your user to login | 04:04 |
| systemdlete | well, a little delay, but then lightdm appears | 04:04 |
| tuxd3v | make abkackup of your config, and experiment with this one :) | 04:04 |
| systemdlete | I will try again, to see if it is faithful | 04:04 |
| systemdlete | tuxd3v: thanks. Let me finish this test with lightdm, then I'll try your config. | 04:05 |
| tuxd3v | no problem | 04:05 |
| systemdlete | came right up again (lightdm) | 04:06 |
| systemdlete | one more time, just for certainty... | 04:07 |
| tuxd3v | systemdlete, solved? | 04:16 |
| systemdlete | hold on | 04:16 |
| systemdlete | still getting slim back | 04:16 |
| tuxd3v | ok | 04:16 |
| systemdlete | I will need desktop to copy in your file -- I hate typing | 04:17 |
| systemdlete | It's coming... it's in its usual 2:10 pause | 04:17 |
| systemdlete | so it is definitely a slim-related issue somehow | 04:17 |
| systemdlete | because lightdm came right up upon boot | 04:17 |
| tuxd3v | activate openssh-server then do a scp to inside of the vm ;) | 04:19 |
| systemdlete | nah | 04:20 |
| systemdlete | this was quick enough | 04:20 |
| systemdlete | rebooting with YOUR slim.conf (I saved the one I had though) | 04:21 |
| systemdlete | not looking too good | 04:21 |
| systemdlete | it's hanging there, blank screen again | 04:21 |
| tuxd3v | humm | 04:22 |
| systemdlete | slim IS running, btw, during this pause | 04:23 |
| systemdlete | I can see it by opening tty1 as root | 04:23 |
| tuxd3v | ls -l /var/run/slim.lock | 04:23 |
| tuxd3v | check if the lock file is there | 04:23 |
| systemdlete | yeppers | 04:23 |
| systemdlete | oh, hold on | 04:24 |
| systemdlete | slim had already come up | 04:24 |
| systemdlete | let me try that again | 04:24 |
| gnarface | systemdlete: when slim is hanging or right after it comes up, can you ssh into the VM to see if anything weird shows up /var/log/Xorg.0.log at that point or right before hand? | 04:24 |
| systemdlete | yeah, I've got tty1 open on root now | 04:24 |
| systemdlete | sure, this reboot | 04:24 |
| gnarface | systemdlete: (sorry i didn't answer the earlier question; i was afk but i would have suggested xdm. still, it looks like lightdm still allowed you to isolate the issue. now i'm wondering if it's related to the graphical drivers somehow) | 04:25 |
| systemdlete | it is, per many entries from google | 04:25 |
| gnarface | inside the VM you should have a lot of control over that though | 04:26 |
| systemdlete | the lock is there, tuxd3v | 04:26 |
| tuxd3v | so its running.. | 04:26 |
| gnarface | you should be able to tell the VM to lie about the hardware, which should allow you any choice of drivers... | 04:27 |
| systemdlete | x is not running | 04:27 |
| systemdlete | now x is running | 04:27 |
| systemdlete | but slim has not appeared yet | 04:27 |
| systemdlete | suspended AIGLX clients for switch... | 04:28 |
| systemdlete | and slim is up | 04:28 |
| tuxd3v | my file has a 'default_user tuxd3v' | 04:30 |
| tuxd3v | that you need to chamge acordingly with your user.. | 04:30 |
| systemdlete | well, the issue seems to be that X is not running and slim is waiting for X to accept connections. | 04:30 |
| tuxd3v | chamge -> change | 04:30 |
| systemdlete | maybe how slim is launching X? | 04:30 |
| gnarface | systemdlete: out of curiosity, how much RAM are you allocating to this VM? it's qemu-kvm? or vbox? | 04:31 |
| systemdlete | vbox 4G | 04:31 |
| systemdlete | and I also checked to see there is ample disk space, swap, etc | 04:32 |
| gnarface | hmm. 4G should be fine | 04:32 |
| systemdlete | yeah | 04:32 |
| systemdlete | Especially merely upon boot... | 04:32 |
| gnarface | i was just thinking maybe if you were assigning it 8MB, Xorg could be hanging just to allocate swap for the video drivers | 04:32 |
| systemdlete | if a boot process ate that much ram, we'd really have issues | 04:32 |
| systemdlete | oh | 04:32 |
| systemdlete | true | 04:32 |
| gnarface | yea you should be fine with as little as 64MB of ram though | 04:33 |
| gnarface | with most drivers anyway | 04:33 |
| systemdlete | the thing is though... let's remember that lightdm has no such issues I can tell | 04:33 |
| gnarface | yea, but not all graphical interfaces are created equally | 04:33 |
| systemdlete | There were several hits when I googled it indicating issues with the drivers. | 04:33 |
| systemdlete | (I agree, which is why I'm tempted to dump slim at this point) | 04:33 |
| gnarface | if for example slim was compositing by default and lightdm was not, then there would be a massive difference in hardware and driver usage just at start | 04:34 |
| gnarface | but, that might mean easy fixes include switching video drivers too | 04:34 |
| systemdlete | I did NOT try to log in because I'd have to install the elogind ware and remove the policykit | 04:34 |
| tuxd3v | verify /var/log/Xorg.0.log, for anny error | 04:34 |
| systemdlete | consolekit I mean | 04:34 |
| systemdlete | http://paste.debian.net/1117981/ | 04:35 |
| gnarface | [ 583.635] (EE) No surface to present from. ??????????????????????? | 04:36 |
| gnarface | never seen this, but it looks like the source can only be a xorg.conf mistake or a bad driver | 04:36 |
| gnarface | also you said it's vbox but this Xorg log is clearly struggling with vmware drivers.... | 04:37 |
| systemdlete | Yeah, trying to recall... | 04:38 |
| systemdlete | I think for vbox users, they call in vmware drivers for some things. But I am not clear on that. | 04:38 |
| systemdlete | check this out: https://forums.gentoo.org/viewtopic-p-7700690.html See last comment | 04:39 |
| gnarface | i recall that vmware had heavy graphical performance issues if you did not install the vmware guest additions package in the VM itself | 04:39 |
| systemdlete | well, I've had the vbox GAs installed there for some time now. | 04:39 |
| gnarface | there might have been permissions settings necessary on the host too | 04:39 |
| systemdlete | dk; dc. I only deal with vbox. If there's any use of vmware it is not anything I did intentionally. | 04:40 |
| gnarface | yea it could be solely a slim bug | 04:40 |
| gnarface | but i don't see any conclusive proof one way or another | 04:40 |
| systemdlete | seems like, yes. Did you see that last comment on that page? | 04:40 |
| systemdlete | no, not conclusive | 04:40 |
| gnarface | i've never heard of it being a thing that vbox requires vmware drivers... that seems wrong to me, but i have not used either so i don't know for sure | 04:41 |
| gnarface | if missing vmware drivers causes performance issues in vmware, and if it's loading vmware drivers and it's *not* vmware, that seems like it could be a source of performance issues as well... | 04:42 |
| gnarface | though honestly i wouldn't even expect it to work at that point, so maybe make double sure this is not actually a vmware vm.... you could be following the wrong debugging process entirely.... | 04:42 |
| systemdlete | it's Vbox. I am 100% sure of that. | 04:44 |
| systemdlete | more likely misconfiguration of X somehow. But, look, this works. I can log in to my desktop, even after the long pause from slim | 04:44 |
| systemdlete | things work; I haven't noticed any problems operationally. I can run my desktop all my fave programs, etc | 04:45 |
| systemdlete | at least until yesterday when this started happening. | 04:45 |
| tuxd3v | It could be on activating 3d aceleration.. a problem with the drivers.. | 04:45 |
| systemdlete | I appreciate the suggestion, of course. | 04:45 |
| systemdlete | I have 3d enabled, along with vt-x and nested pages | 04:45 |
| gnarface | well misconfiguration of X can include loading the wrong drivers... can you think of any Xorg related changes to the config or drivers before the last time you restarted it, however long it was before that? | 04:45 |
| tuxd3v | gnarface, you have eagle eyes.. you catched the bug very fast.. | 04:45 |
| systemdlete | caught what bug very fast? | 04:46 |
| gnarface | tuxd3v: grep skills ;) | 04:46 |
| tuxd3v | your bug | 04:46 |
| systemdlete | what bug is it? | 04:46 |
| tuxd3v | gnarface, hehe | 04:46 |
| tuxd3v | its a bug related with xorg, or drivers of virtual box | 04:47 |
| tuxd3v | I saw a lot of entries about that.. | 04:47 |
| systemdlete | well, perhaps. | 04:47 |
| systemdlete | Or it could be slim. | 04:47 |
| tuxd3v | some even states that it could be when activating 3d aceleration.. | 04:47 |
| systemdlete | Maybe it doesn't play well in all configurations of X. I mean, it generates errors saying it cannot get X to stop | 04:47 |
| gnarface | well, it could be a slim bug, but it could also just be slim compositing by default that exposes a driver bug or a mesa bug.... | 04:48 |
| systemdlete | well, if so, it is not a problem for lightdm! | 04:48 |
| systemdlete | gnarface: +1 absolutely | 04:48 |
| systemdlete | something in its environment does not like what it is trying to do, even though until just days ago, everything seemed to be fine | 04:48 |
| systemdlete | gnarface: the only updates I recall would be a few days ago when I did that last update/upgrade | 04:49 |
| systemdlete | I think it included the -11 kernel, but I'm not 100% sure | 04:49 |
| systemdlete | but I don't remember this long pause until slim presented the login | 04:49 |
| tuxd3v | systemdlete, truncate the log file and reboot | 04:50 |
| gnarface | systemdlete: well, the kernel change would have changed the video driver version at the very least. that is a conceivable cause of regression, but a mesa update could easily have accompanied it | 04:50 |
| tuxd3v | > /var/log/Xorg.0.log | 04:50 |
| tuxd3v | then paste the log again, if you can | 04:50 |
| tuxd3v | becasue it will be clean | 04:50 |
| tuxd3v | a clean bootup | 04:50 |
| tuxd3v | I suspect that somehow it fallback to modesetting.. | 04:51 |
| systemdlete | know what else? I tried nomodeset. I recall reading something about new configurations that pass the resolution along to the kernel (KMS?) but it seems this does not always work for everyone, so they have to disable kernel modesetting and let X do it | 04:51 |
| tuxd3v | after a minute or so.. | 04:51 |
| systemdlete | And the first time, modeset did seem to clear up the problem, the same way that fsmithred's suggestion did (replacing that driver). But after that, the same problem kept appearing. | 04:52 |
| gnarface | that part makes me wonder if changes to your VMs are even being saved | 04:52 |
| systemdlete | I even tried nouveaux.modeset=0 on a lark because someone suggested it on one of the sites | 04:52 |
| systemdlete | Are you referring to disk images? or maybe the vbox config's? | 04:53 |
| systemdlete | If so, then I should be able to make some changes to a few files, reboot and see if the changes are still there. | 04:53 |
| gnarface | well, i'm non-speficially referring to both | 04:53 |
| gnarface | but i primarily had the disk images themselves in mind | 04:53 |
| systemdlete | hmmm. | 04:53 |
| systemdlete | file perms... hmmmmm | 04:54 |
| gnarface | since a reasonably common configuration is read-only images that spawn disposably to ram, and don't get saved | 04:54 |
| gnarface | just make sure you're not doing something like that every time you make a change that disappears the next time you reboot the VM | 04:54 |
| systemdlete | file perms look fine | 04:54 |
| gnarface | i mean it would be obvious if they weren't being saved. the timestamps on them wouldn't update when the VM shuts down | 04:55 |
| systemdlete | I very much doubt that. | 04:55 |
| systemdlete | timestamps look up to date | 04:55 |
| systemdlete | virtual disk images have current time on them | 04:55 |
| systemdlete | meta data can only be changed when VM is quiescent (mostly, except for a few fields ones I rarely change) | 04:56 |
| systemdlete | There was one issue, though, that I am still pursuing but on my testbox | 04:57 |
| systemdlete | Listen to this, gnarface: A few days ago, I went to open a url I received in an email -- from a very reputable source, though I cannot say the same for their IT work, I simply dk -- and it began loading, the display flashed (when in the full screen mode) and the web page froze, the time in the VM froze, and I could not even break out of it. | 04:58 |
| systemdlete | I could ssh in from another system, so the host was actually OK (I think). | 04:59 |
| systemdlete | I repeated that a couple times before I got fed up of rebooting the host. | 04:59 |
| systemdlete | I am going to see if I can replicate that scenario in a VM on my testbox (different physical machine) | 04:59 |
| systemdlete | btw, when I open the same link in a browser on the host, the page hung, but no impact on the rest of the system. | 05:00 |
| gnarface | hmmmmm.... does sound suspicious | 05:00 |
| systemdlete | yeah. | 05:00 |
| tuxd3v | are you using nouveau driver ? | 05:01 |
| gnarface | cross-platform graphical vulnerabilities have been identified in the nvidia official driver before. it's a risk for anything using direct rendering though | 05:01 |
| systemdlete | If anything, I'm more suspicious of that incident than the current one. The current one could be another manifestation of the same issue, but it's hardly deadly like the complete lockup one. | 05:01 |
| systemdlete | this is an AMD system. Not sure on the hw specs atm | 05:02 |
| systemdlete | Its' an ASUS board | 05:02 |
| systemdlete | M5A78L_M_USB3 | 05:02 |
| systemdlete | I'm using the onboard video | 05:03 |
| tuxd3v | you have amdgpu enabled or you are using another driver? | 05:03 |
| systemdlete | If I can't replicate that scenario on the other PC, maybe I'll install a spare video card, though it's kind of old | 05:03 |
| systemdlete | coach me on that, and I'll tell you... | 05:04 |
| systemdlete | (I'm no hw guru) | 05:04 |
| systemdlete | gnarface: Sorry I didn't investigate the HW/nvidia side first. | 05:04 |
| systemdlete | Just didn't think of it. I was thinking of it all as VM and software. I don't typically look for HW problems. | 05:05 |
| systemdlete | (though I've certainly had some over the many years) | 05:05 |
| tuxd3v | xserver-xorg-video-radeon | 05:06 |
| tuxd3v | this is what you should be using.. I mean the package | 05:06 |
| tuxd3v | lspci|grep -i AMD | 05:07 |
| systemdlete | it is def installed | 05:07 |
| tuxd3v | 'Integrated ATI Radeon™ HD 3000 GPU' | 05:07 |
| tuxd3v | I don't know if it is supported by amdgpu, I believe that you need to be with radeon driver.. | 05:08 |
| systemdlete | it's installed, but does not show in lspci: http://paste.debian.net/1117984/ | 05:09 |
| systemdlete | this board has been around a number of years. My impression is that it is fairly popular. So I'd think, by now, there would be well-established support for its hardware configuration. | 05:10 |
| systemdlete | but I could be wrong about it. | 05:11 |
| tuxd3v | I mean run it in the host os | 05:14 |
| tuxd3v | :) | 05:14 |
| tuxd3v | you are in vmware? | 05:14 |
| systemdlete | oh | 05:14 |
| tuxd3v | 00:02.0 VGA compatible controller: VMware SVGA II Adapter | 05:14 |
| systemdlete | no virtualbox | 05:14 |
| systemdlete | 01:05.0 VGA compatible controller: Advanced Micro Devices, Inc. [AMD/ATI] RS780L [Radeon 3000] | 05:15 |
| systemdlete | (from host) | 05:15 |
| tuxd3v | yeah | 05:16 |
| tuxd3v | I don know if it has support for amdgpu | 05:16 |
| tuxd3v | do a 'lsmod' in the host | 05:16 |
| systemdlete | gnarface, tuxd3v: I recall reading a post about how some people in linux do not like the quality of vbox's video drivers. That might be why they configure vbox linux systems with the vmware drivers | 05:16 |
| systemdlete | lsmod |grep gpu shows nada | 05:17 |
| tuxd3v | lsmod|grep -Ei "amd|radeon" --color | 05:18 |
| systemdlete | amdkfd, radeon, kvm_amd... yes they are there | 05:18 |
| systemdlete | http://paste.debian.net/1117985/ | 05:19 |
| tuxd3v | yup | 05:21 |
| tuxd3v | you are using the admkfd, that uses the userspace radeon driver | 05:21 |
| systemdlete | moreover, see: https://forums.virtualbox.org/viewtopic.php?t=91874&start=15 | 05:21 |
| tuxd3v | not the amdgpu one.. | 05:21 |
| systemdlete | " if i purge xserver-xorg-video-vmware, my debian sid vm can't start x, because modesetting-ddx does not work (error: failed to add fb -22)." | 05:21 |
| systemdlete | so it sounds like that vmware driver is essential | 05:22 |
| systemdlete | s/is/might be/ | 05:22 |
| systemdlete | now I've got mouse hesitation... on my host. Probably because of the cheap kvm switch. | 05:26 |
| systemdlete | I've had periodic issues with that POS too. | 05:26 |
| systemdlete | tuxd3v: So what should I do? | 05:33 |
| systemdlete | or do I need to do anything? | 05:35 |
| tuxd3v | don't know, but one thing that I would try, would be to disable 3d acceleration in the vm | 05:36 |
| tuxd3v | amd see if that solves | 05:36 |
| systemdlete | ok | 05:36 |
| tuxd3v | also check the Xorg logs in the host | 05:36 |
| systemdlete | gimme a moment | 05:36 |
| tuxd3v | to see if my config is ok | 05:36 |
| systemdlete | disabling 3d did nothing to change the problem | 05:42 |
| systemdlete | still hangs upon boot, sorry. | 05:42 |
| systemdlete | last 1000 lines of host xorg log: http://paste.debian.net/1117989/ | 05:44 |
| systemdlete | (seems to be a limit) | 05:44 |
| gnarface | does it dump those repeated modeline blocks when a VM starts up? | 05:47 |
| systemdlete | on the host? | 05:47 |
| gnarface | yea, that last paste was from the host, right? | 05:48 |
| systemdlete | yes | 05:48 |
| gnarface | it's weird that it's dumping all the detected modelines a half dozen times in a row. does it dump once for each VM you start, or is it actually dumping that block of text 6x times (or whatever it is) every time you start just 1 VM? | 05:48 |
| systemdlete | there is one way to find out... hold on | 05:49 |
| tuxd3v | yeah too many modelines several times | 05:50 |
| gnarface | i'm hypothesizing the delay being related to some sort of auto-detect loop now | 05:50 |
| gnarface | i didn't figure that would be the case for a VM but maybe in fullscreen mode ... | 05:50 |
| gnarface | i dunno | 05:50 |
| gnarface | seems suspicious though in light of the other misbehaviors | 05:51 |
| tuxd3v | gnarface, yes it could be | 05:51 |
| gnarface | it could be nothing too. i'm not using any radeon stuff here. | 05:52 |
| systemdlete | ok, it looks like those logs were because of yours truly resetting my kvm switch. | 05:54 |
| systemdlete | the problem there, btw, was a weak battery. I replaced it and seems like things are good again. | 05:54 |
| systemdlete | (for now) | 05:54 |
| systemdlete | each time I reset or change to a different pc, I was causing that flood of messages in the x.org log | 05:55 |
| systemdlete | interestingly, though, not one single message upon firing up or shutting down a VM! | 05:55 |
| systemdlete | Even when the VM changes its video, there are no new messages on the host in the log | 05:55 |
| systemdlete | that could have to do with the verbosity the x log is set to | 05:56 |
| systemdlete | well, actually, as regards fullscreen mode, I've been doing all these tests for you in windowed mode, not full screen. | 05:57 |
| systemdlete | Full screen mode: Same thing. | 05:58 |
| systemdlete | Hey, I just had this nifty idea. | 06:06 |
| systemdlete | I will stop slim, then launch it in strace, capturing the output for a few minutes. | 06:06 |
| systemdlete | :D | 06:06 |
| systemdlete | see what it is "Thinking" | 06:06 |
| golinux | systemdlete: Are you still slimming around? | 06:15 |
| systemdlete | trying to, yeah... | 06:15 |
| systemdlete | or did you mean "sliming" around? lol | 06:16 |
| golinux | That's a real marathon . . . | 06:16 |
| golinux | I've never had an issue with slim but some folks do. | 06:16 |
| golinux | Actually I thought of slumming | 06:16 |
| golinux | So many OT possibilities. | 06:17 |
| systemdlete | well, I went to the doc today. I am hitting 180 lbs. that's enough | 06:17 |
| systemdlete | yeah, keep it up golinux and get your self kicked by the mods! | 06:17 |
| systemdlete | :p | 06:17 |
| golinux | LOL! | 06:17 |
| systemdlete | you are a primary offender that way around here... | 06:18 |
| systemdlete | though I'm probably right behind you... like #2 or 3 | 06:18 |
| systemdlete | right now, I'm trying to strace the damned thing. | 06:21 |
| systemdlete | Hmmm. Looks like slim is epoll_wait'ing on /proc/filesystems, but I'm not sure if I'm seeing this right. | 06:34 |
| systemdlete | nope. Its waiting on slim.lock | 06:37 |
| systemdlete | the last thing that process opens looks like /proc/filesystems, then it does that poll | 06:45 |
| systemdlete | (I think, still looking at this) | 06:45 |
| systemdlete | One thing is for certain: slim is waiting (epoll'ing) for some time, waiting on one or more file descriptors. | 07:03 |
| systemdlete | bbl. Food! | 07:05 |
| TwistedFate | what is devuan using for config file management? | 07:15 |
| TwistedFate | i removed and purged a package, now the config files will not install when i re-install it.. | 07:16 |
| gnarface | systemdlete: maybe this is a simple issue with a default setting in /etc/security/limits.conf being starved out by the VMs... i used to have to raise nofile for tf2 servers... | 07:16 |
| systemdlete | no files is 1024. I *think* that's plenty... | 07:20 |
| systemdlete | I'm now trying to ascertain WHICH files epoll_wait is waiting on. | 07:21 |
| systemdlete | prob easier to examine the source at this point... | 07:21 |
| TwistedFate | guys, how do i force apt or dpkg to reinstall the package configuration as well? | 07:25 |
| gnarface | TwistedFate: it's supposed to do that normally ... | 07:26 |
| TwistedFate | i removed a package, purged it and deleted the configs directory, now when i reinstall it, it doesn't install configs too :/ | 07:26 |
| gnarface | systemdlete: the new default is 8192 i think (what i had to raise it to for the TF2 server) | 07:26 |
| gnarface | TwistedFate: were they in a separate package? | 07:27 |
| systemdlete | when you say "the new default is..." can you explain that please? thanks | 07:27 |
| TwistedFate | gnarface: i'm unsure, i installled nginx and it gave me /etc/nginx and bunch of other directories and configs in that directory | 07:27 |
| TwistedFate | now they're gone | 07:28 |
| gnarface | systemdlete: well it's not that new of a default, but you said it was 1024 which i recall having not been the default for many years now | 07:28 |
| gnarface | TwistedFate: you sure you didn't manually copy them from /usr/share/doc/nginx? | 07:29 |
| TwistedFate | gnarface: yes, i'm sure, i just manually deleted them | 07:29 |
| systemdlete | gnarface: So, did I miss an update? Or just some brain cells? | 07:31 |
| * systemdlete really needs to start smoking... | 07:32 | |
| gnarface | systemdlete: well probably what happened is you kept your old config file and didn't merge any changes during an update (expected behavior) or else maybe your kernel is super old... but you said this was ascii, right? so it shouldn't be that old | 07:32 |
| systemdlete | 4.9.0-11, yes sir | 07:33 |
| systemdlete | are you telling me yours is set to 4192, but I somehow missed that change? | 07:33 |
| systemdlete | 8192 | 07:34 |
| gnarface | well the first part i'm sure of. the second part ... only 80-90% | 07:34 |
| gnarface | i know for sure i had to increase it for the TF2 server. i think i also had to increase it for something in wine at one point too. but i'm less sure that i've heard the default was changed to that later. | 07:35 |
| gnarface | it is something easy enough to test though | 07:35 |
| systemdlete | I'm having some wine myself right now | 07:36 |
| gnarface | it looks like that config file is part of libpam-modules, maybe verify that you have the ascii version of that just as a sanity check | 07:36 |
| systemdlete | sure, I could. But I think I want to try the getting rid of all things ~bpo first | 07:36 |
| TwistedFate | how can i check apt/apt-get/aptitude commands history and output? | 07:45 |
| systemdlete | gnarface: Didn't you say that ~bpo packages are from backports? | 07:54 |
| systemdlete | I've still got them even though I've cleared the packages cache (archive in /var) and commented them out in sources | 07:55 |
| gnarface | systemdlete: i did say that, but then shortly afterwards fsmithred said something about seeing them in ascii-security for some reason... | 07:55 |
| gnarface | systemdlete: which ones are you seeing still, specifically? | 07:56 |
| systemdlete | policykit | 07:56 |
| systemdlete | that's ok | 07:56 |
| gnarface | hmmm, that's a package they forked so maybe that is on purpose. any others? | 07:56 |
| systemdlete | well, I meant all the policy-kit related stuff | 07:57 |
| gnarface | TwistedFate: check in /var/log/apt/ | 07:57 |
| gnarface | TwistedFate: aptitude may be separate from that | 07:58 |
| TwistedFate | 'apt-get remove nginx && apt-get purge nginx' shows in my 'history |grep purge' but it's not showing in /var/log/apt/history.log for some reason.. | 08:12 |
| gnarface | systemdlete: there's some way to check if it's hitting the limits.conf limit, but i forget exactly how... you might need to enable some debugging in the logs | 08:12 |
| gnarface | TwistedFate: as i understand it, you only purge installed packages.... so it would have probably just ignored the second call or errored it out after succeeding at a regular uninstall of nginx. | 08:13 |
| Jjp137 | no you can still purge packages if they left configuration files behind | 08:23 |
| Jjp137 | although purge also does remove so you don't need to do both commands | 08:24 |
| TwistedFate | welp | 09:08 |
| TwistedFate | turns out it didn't want to reinstall the configs because of an additional file/package that needed to be purged also | 09:08 |
| systemdlete | gnarface: Is this pursuit -- to determine the cause of the long hang in slim -- a worthy goal for the furtherment of science and human knowledge? Or could I just install lightdm and be done with it? | 10:21 |
| gnarface | systemdlete: oh i thought you had some specific reason to want slim. like i said, i personally would have long ago given up on graphical login managers entirely for something like this, or switched to something else | 10:23 |
| gnarface | systemdlete: if you don't care, what you should do is file a bug report and move on | 10:23 |
| systemdlete | Many already have. Or at least, this is a well-known issue, per google. But, yes, I suppose I should do that favor and file it. | 10:24 |
| systemdlete | But where? Devuan? Debian? X.org? slim? | 10:24 |
| gnarface | well i'd first check to see if one is already filed with debian | 10:24 |
| systemdlete | And given that I don't know when the issue arose precisely... maybe I would just be adding more clutter to somebody's bug list. | 10:25 |
| systemdlete | debian. OK | 10:25 |
| r3boot | systemdlete: did you check ~/.xsession-error to see if there's something that's blocking the startup? | 10:25 |
| systemdlete | oooh! | 10:25 |
| systemdlete | nice one. NO, I hadn't. | 10:25 |
| systemdlete | but, then again, would this show up there? This is in the display manager, so I would think .xsession-error would not even have been opened yet? | 10:26 |
| systemdlete | We've been looking at the X org log files in /var/log | 10:26 |
| r3boot | I am missing backlog, so I might have missed something; Avid slim user here (on arch tho) | 10:26 |
| systemdlete | np | 10:27 |
| gnarface | r3boot: he's using it in a VM and it delays startup significantly compared to lightdm | 10:27 |
| systemdlete | I think DM's, generally, are user-agnostic until the point a login is successful. | 10:27 |
| gnarface | r3boot: (the short version) | 10:27 |
| systemdlete | gnarface: (thank you) | 10:28 |
| r3boot | gnarface: check, thnx | 10:28 |
| systemdlete | of course, I'll also have to replace the policyshit packages with elogind | 10:31 |
| systemdlete | why does the screen flash just as the lightdm login appears? I noticed that with slim also. | 10:42 |
| systemdlete | another question: fwanalog ran for the first time in the same VM as we were looking at w/r/t slim. However, I don't remember actually installing it, and neither does dpkg.* or any of the files in /var/log/apt. | 11:52 |
| systemdlete | I DO recall *considering* it for helping to track down some mysterious packets. | 11:52 |
| gnarface | hmm, not sure about that | 11:53 |
| systemdlete | I *may* indeed have installed it. But I would think it would be recorded somewhere. | 11:53 |
| gnarface | what about /var/log/dpkg.log ? | 11:54 |
| systemdlete | see ^^^ | 11:54 |
| systemdlete | no, nothing there either | 11:54 |
| gnarface | interesting | 11:54 |
| systemdlete | I did: zgrep fwanalog /var/log/dpkg.* /var/log/apt/* | 11:54 |
| gnarface | is there a /var/log/aptitude? | 11:55 |
| systemdlete | hmmm. | 11:55 |
| systemdlete | no | 11:56 |
| gnarface | does it show up in dpkg -l ? | 11:57 |
| systemdlete | yes | 11:57 |
| gnarface | already there with the base system maybe? | 11:58 |
| systemdlete | no way. I know that for a fact. It was not there previously (yesterday). If it had been I'd been getting reports daily. This is the first time. | 11:58 |
| systemdlete | Like I said, I was considering installing it to see if it could help me get more info on some packets being logged. | 11:59 |
| gnarface | interesting | 11:59 |
| systemdlete | But I don't recall if I actually installed it or not last night -- I was having so much fun with these other issues. | 11:59 |
| systemdlete | omg. | 12:02 |
| systemdlete | I just re-installed gufw and it installs the policykit crap again | 12:02 |
| systemdlete | so now I've got elogind AND some policykit stuff also. | 12:02 |
| systemdlete | This is just too much fun. | 12:03 |
| systemdlete | System (VM) is still bootable, login-able, desktop-able | 12:07 |
| systemdlete | So I guess it's OK, even with both. | 12:08 |
| systemdlete | seriously, I am really ready to just rebuild this whole mess from scratch and restore my user files | 12:10 |
| systemdlete | I realize that's the sleazy way out of this, but these package dependencies are driving me batty | 12:10 |
| systemdlete | gufw depends on policykit prob because it needs to authenticate the user running it. | 12:11 |
| systemdlete | (but it must be welded on to the method this way?) | 12:11 |
| systemdlete | enough for one night. bbl -- thanks to all who helped. | 12:13 |
| gnarface | apparently, but systemd could have something to do with it, i don't know. | 12:13 |
| systemdlete | thx, gnarface. Time for sleep now. | 12:13 |
| gnarface | peace | 12:13 |
| systemdlete | same | 12:13 |
| dbristow | Sooo close to finishing downloading the 2.1 torrent. I think I am still lacking about 10M of the files. | 21:55 |
| dbristow | 99.9% done | 21:56 |
| dbristow | 23.10G out of 23.11G | 21:56 |
Generated by irclog2html.py 2.17.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!