00:04 | vagrantc has left IRC (vagrantc!~vagrant@2600:3c01:e000:21:21:21:0:100b, Quit: leaving) | |
02:32 | eu^3sub-174-250- has joined IRC (eu^3sub-174-250-!~eu^3sub-1@2600:1000:bf0a:157d:b8f3:5cb9:7e5a:5d91) | |
02:35 | <eu^3sub-174-250-> Hello, A new server build of ltsp on Debian Bullseye . Using the same two thin clients, Dell All-in-One machines about 12 years old. They bothe will not shut down they reboot. Even trying to manually shut each one off,they restart . These have worked fine for about three previous Debian/Ltsp installes.
| |
02:35 | fottsia[m] has joined IRC (fottsia[m]!~fottsiama@2001:470:69fc:105::48bb) | |
03:42 | eu^3sub-174-250- has left IRC (eu^3sub-174-250-!~eu^3sub-1@2600:1000:bf0a:157d:b8f3:5cb9:7e5a:5d91, Quit: Client closed) | |
04:27 | alkisg_web has joined IRC (alkisg_web!~alkisg_we@srv1-dide.ioa.sch.gr) | |
05:50 | ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz) | |
06:04 | alkisg_web has left IRC (alkisg_web!~alkisg_we@srv1-dide.ioa.sch.gr, Quit: Ping timeout (120 seconds)) | |
06:19 | woernie has left IRC (woernie!~werner@p5ddec1ab.dip0.t-ipconnect.de, Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) | |
06:35 | alkisg_web has joined IRC (alkisg_web!~alkisg_we@srv1-dide.ioa.sch.gr) | |
07:23 | Rohit has joined IRC (Rohit!~Rohit@117.248.249.248) | |
07:24 | Rohit has left IRC (Rohit!~Rohit@117.248.249.248, Client Quit) | |
07:25 | decenttp has joined IRC (decenttp!~decenttp@117.248.249.248) | |
07:31 | <decenttp> alkisg: this is regarding the thread https://github.com/ltsp/ltsp/discussions/560
| |
07:31 | <alkisg> Hey decenttp . If you could please share your server screen with me, we could check it out quickly, as I'm a bit time-pressed currently. See the bot instructions below:
| |
07:31 | !vnc-edide
| |
07:31 | <ltspbot> vnc-edide: To share your screen with me, open Epoptes → Help menu → Remote support → Host: srv1-dide.ioa.sch.gr, and click the Connect button
| |
07:33 | <decenttp> hi alkisg: i just connected on Epoptes
| |
07:43 | <alkisg> decenttp: your nfs server share is read-only, and you didn't have nfs-common in the image
| |
07:44 | <decenttp> alkisg: so i would need to re-install ltsp server?
| |
07:44 | <alkisg> If I restart the nfs server to load the new share, it may block your users for up to a minute
| |
07:44 | It's better to test this when there are no users logged in
| |
07:44 | <decenttp> alkisg: that is fine
| |
07:44 | <alkisg> No of course not. Just `sudo systemctl restart nfs-server`
| |
07:44 | OK then
| |
07:49 | decenttp: something is wrong there that I can't pinpoint, and the nfs share ends up read-only,
| |
07:49 | but I see that you have local disks there
| |
07:49 | Just make a swap partition and ltsp will use it
| |
07:49 | See if that makes things more stable
| |
07:49 | Do you want me to launch gparted and do that now?
| |
07:49 | For this specific user?
| |
07:49 | <decenttp> that's fine
| |
07:50 | alkisg: that is fine we can do that. with current issue they are not able to use it anyways.
| |
07:50 | <alkisg> I'm not sure if that means that YOU'll do it later on, or if I should do it now
| |
07:50 | Should I do it now?
| |
07:50 | <decenttp> alkisg: if you can do it now it would be great
| |
07:51 | <alkisg> OK
| |
07:51 | <decenttp> alkisg: i am unable to pin point at a specific issue. we did not have this issue until we upgraded
| |
07:51 | <alkisg> Tell the user to pause for a bit :D
| |
07:51 | <decenttp> ok
| |
07:52 | alkisg: stopped
| |
07:53 | <alkisg> So just do the same in other computers and reboot them, ltsp will automatically use the swap partition
| |
07:55 | <decenttp> alkisg: ok, you were quick. it happened when i went to the users desk
| |
07:56 | alkisg: but i understand what you want me to do. I would try to figure it out and do it on other clients.
| |
07:57 | <alkisg> Swap: 8395772 0 8395772
| |
07:57 | That's the key line there
| |
07:57 | <decenttp> ok
| |
07:57 | <alkisg> OK, closing, let's talk again later if this persists
| |
07:58 | <decenttp> alkisg: ok thanks a ton for your time. I would implement on other clients and report back.
| |
08:00 | <alkisg> 👍️
| |
09:00 | fottsia[m] has left IRC (fottsia[m]!~fottsiama@2001:470:69fc:105::48bb, Quit: You have been kicked for being idle) | |
09:53 | woernie has joined IRC (woernie!~werner@p578bb7b6.dip0.t-ipconnect.de) | |
10:21 | eu^94-183-177-68 has joined IRC (eu^94-183-177-68!~eu^94-183@94.183.177.68) | |
10:26 | eu^94-183-177-68 has left IRC (eu^94-183-177-68!~eu^94-183@94.183.177.68, Ping timeout: 256 seconds) | |
10:55 | woernie has left IRC (woernie!~werner@p578bb7b6.dip0.t-ipconnect.de, Ping timeout: 265 seconds) | |
12:25 | <MUHWALT> hey alkisg, odd question: clients running chrome seem to write to their home dir _a lot_. One of my LTSP5 installs shares a RAID10 array with a few other VMs and IO seems to be an issue sometimes. Any experience calming google chrome down, or have some magic to mitigate clients writing a lot?
| |
12:25 | fottsia[m] has joined IRC (fottsia[m]!~fottsiama@2001:470:69fc:105::48bb) | |
12:25 | <MUHWALT> I'm assuming it's local cacheing, e.g., to ~/.cache/google-chrome/*, but haven't actually confirmed
| |
12:26 | standard ssh home mounting
| |
12:27 | (obligatory SOON on upgrading to new ltsp ;) )
| |
13:20 | fottsia[m] has left IRC (fottsia[m]!~fottsiama@2001:470:69fc:105::48bb, Remote host closed the connection) | |
13:20 | ServerStatsDisco has left IRC (ServerStatsDisco!~serversta@2001:470:69fc:105::1a, Read error: Connection reset by peer) | |
13:20 | zelt1954[m] has left IRC (zelt1954[m]!~zelt1954m@2001:470:69fc:105::1:c34, Read error: Connection reset by peer) | |
13:20 | renots[m] has left IRC (renots[m]!~renotsmat@2001:470:69fc:105::f879, Remote host closed the connection) | |
13:20 | ABIXadamjurkiewi has left IRC (ABIXadamjurkiewi!~adamjurki@2001:470:69fc:105::f1a6, Remote host closed the connection) | |
13:20 | sfxworks[m] has left IRC (sfxworks[m]!~sfxworksm@2001:470:69fc:105::e5d4, Write error: Broken pipe) | |
13:20 | DouglasGiovaniOe has left IRC (DouglasGiovaniOe!~doguibnum@2001:470:69fc:105::1:929, Write error: Connection reset by peer) | |
13:20 | Jonny[m] has left IRC (Jonny[m]!~jonnymatr@2001:470:69fc:105::b958, Write error: Connection reset by peer) | |
13:20 | BrunoBinet[m] has left IRC (BrunoBinet[m]!~bbinetmat@2001:470:69fc:105::68c6, Write error: Connection reset by peer) | |
13:20 | FotisEAlexakos[m has left IRC (FotisEAlexakos[m!~fotisalex@2001:470:69fc:105::f874, Write error: Connection reset by peer) | |
13:20 | Klampfradler has left IRC (Klampfradler!~nikmatrix@2001:470:69fc:105::1f6b, Write error: Connection reset by peer) | |
13:20 | alkisg has left IRC (alkisg!~alkisg@2001:470:69fc:105::2d3, Write error: Connection reset by peer) | |
13:20 | thanosx[m] has left IRC (thanosx[m]!~thanosxma@2001:470:69fc:105::1:323, Read error: Connection reset by peer) | |
13:20 | dnartreb[m] has left IRC (dnartreb[m]!~dnartrebm@2001:470:69fc:105::d6fa, Write error: Connection reset by peer) | |
13:20 | gidarakos[m]1 has left IRC (gidarakos[m]1!~gidarakos@2001:470:69fc:105::ce6, Remote host closed the connection) | |
13:20 | MikeGabriel[m] has left IRC (MikeGabriel[m]!~mikegabri@2001:470:69fc:105::1:90d, Read error: Connection reset by peer) | |
13:21 | ServerStatsDisco has joined IRC (ServerStatsDisco!~serversta@2001:470:69fc:105::1a) | |
13:21 | FotisEAlexakos[m has joined IRC (FotisEAlexakos[m!~fotisalex@2001:470:69fc:105::f874) | |
13:23 | alkisg has joined IRC (alkisg!~alkisg@2001:470:69fc:105::2d3) | |
13:23 | dnartreb[m] has joined IRC (dnartreb[m]!~dnartrebm@2001:470:69fc:105::d6fa) | |
13:23 | gidarakos[m] has joined IRC (gidarakos[m]!~gidarakos@2001:470:69fc:105::ce6) | |
13:23 | Jonny[m] has joined IRC (Jonny[m]!~jonnymatr@2001:470:69fc:105::b958) | |
13:23 | sfxworks[m] has joined IRC (sfxworks[m]!~sfxworksm@2001:470:69fc:105::e5d4) | |
13:23 | BrunoBinet[m] has joined IRC (BrunoBinet[m]!~bbinetmat@2001:470:69fc:105::68c6) | |
13:23 | thanosx[m] has joined IRC (thanosx[m]!~thanosxma@2001:470:69fc:105::1:323) | |
13:23 | ABIXadamjurkiewi has joined IRC (ABIXadamjurkiewi!~adamjurki@2001:470:69fc:105::f1a6) | |
13:23 | MikeGabriel[m] has joined IRC (MikeGabriel[m]!~mikegabri@2001:470:69fc:105::1:90d) | |
13:23 | renots[m] has joined IRC (renots[m]!~renotsmat@2001:470:69fc:105::f879) | |
13:23 | DouglasGiovaniOe has joined IRC (DouglasGiovaniOe!~doguibnum@2001:470:69fc:105::1:929) | |
13:23 | fottsia[m] has joined IRC (fottsia[m]!~fottsiama@2001:470:69fc:105::48bb) | |
13:23 | zelt1954[m] has joined IRC (zelt1954[m]!~zelt1954m@2001:470:69fc:105::1:c34) | |
13:23 | Klampfradler has joined IRC (Klampfradler!~nikmatrix@2001:470:69fc:105::1f6b) | |
13:27 | ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving) | |
13:49 | ricotz has joined IRC (ricotz!~ricotz@ubuntu/member/ricotz) | |
14:14 | <alkisg> MUHWALT: in my schools we're using firefox, so I don't have a lot of experience with google chrome. On a big non-school site where we're using it, the clients have 16 GB RAM, so we're actually symlinking .cache to /run/user/1000/.cache, to a tmpfs, and we avoid all traffic
| |
14:15 | <MUHWALT> that's helpful, thanks!
| |
14:20 | how do you set that up, in the user's .profile or something?
| |
14:24 | decenttp has left IRC (decenttp!~decenttp@117.248.249.248, Quit: Client closed) | |
14:29 | <alkisg> MUHWALT: we have both a PAM hook and a mate-session override, so we're able to do all kinds of fancy stuff
| |
14:29 | I think this specific one comes from /usr/local/bin/mate-session
| |
14:30 | <MUHWALT> gotcha... interestingly, I installed fatrace on one of the ltsp server and it looks like the majority of writes are to /home/**/.config/google-chrome, not .cache =/
| |
14:31 | <alkisg> In .config? Strange, what does it write in there?! :D
| |
14:31 | <MUHWALT> in any case, keeping browser caching from hitting the server would probably be a good thing!
| |
14:31 | something to the user's profile
| |
14:32 | browser metrics, cookies... strange that those are in .config and not .cache
| |
14:32 | at least the cookies
| |
14:33 | there are definitely large write "storms" to .cache/google-chrome though, so I'm sure there's a benefit to be had redirecting that to local ram
| |
14:34 | our homes are on a 7200k mechanical raid10 array, along with a couple other VMs... that array gets a workout. If I copy a large file over, fatclients start dropping, lol
| |
14:34 | a pain when I need to resize virtual disks!
| |
14:35 | <alkisg> In the big setup we're using zfs with 2*2TB nvme drives, load balanced in 4 identical ltsp servers
| |
14:36 | With 40 gbps networking each, the /home throughput is amazing, more than 10 GB/sec
| |
14:36 | <MUHWALT> Yeah, I'm budgeting a few new servers for next year, which will be 100% SSD
| |
14:36 | we're not maxing out the NICs, I think it's just these "lots of little writes" that are killing the IO performance
| |
14:37 | <alkisg> It might be worth it to explore ceph or rdb sometime, maybe they're optimized to cache the little writes
| |
14:38 | <MUHWALT> I'm *hoping* just throwing fast disks at it will fix it without me needing to do a bunch of research ;)
| |
14:38 | or splitting the disks up into smaller arrays and manually balancing
| |
14:39 | I think this array is just over provisioned, honestly
| |
14:39 | too many VMs fighting for time
| |
14:41 | <alkisg> Sure for 100-200 users plain SSDs should suffice
| |
15:10 | woernie has joined IRC (woernie!~werner@p5ddec1ab.dip0.t-ipconnect.de) | |
16:00 | fottsia[m] has left IRC (fottsia[m]!~fottsiama@2001:470:69fc:105::48bb, Quit: You have been kicked for being idle) | |
17:07 | vagrantc has joined IRC (vagrantc!~vagrant@2600:3c01:e000:21:21:21:0:100b) | |
18:35 | lucascastro has left IRC (lucascastro!~lucascast@177-185-130-117.dynamic.isotelco.net.br, Ping timeout: 245 seconds) | |
20:19 | alkisg_web has left IRC (alkisg_web!~alkisg_we@srv1-dide.ioa.sch.gr, Quit: Client closed) | |
21:47 | ricotz has left IRC (ricotz!~ricotz@ubuntu/member/ricotz, Quit: Leaving) | |