IRC chat logs for #ltsp on irc.libera.chat (webchat)
Channel log from 20 December 2012 (all times are UTC)
| 00:16 | vagrantc has joined IRC (vagrantc!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net) | |
| 00:16 | vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc) | |
| 01:11 | fuzai has joined IRC (fuzai!~fuzai@71-217-108-78.tukw.qwest.net) | |
| 01:12 | F-GT has joined IRC (F-GT!~phantom@ppp59-167-136-109.static.internode.on.net) | |
| 01:18 | dgeary2 has joined IRC (dgeary2!~david@2001:388:e001:1e00:914:40fb:752:d15) | |
| 01:24 | andygraybeal has joined IRC (andygraybeal!~andy@h193.205.130.174.dynamic.ip.windstream.net) | |
| 01:25 | vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving) | |
| 01:28 | fuzai has left IRC (fuzai!~fuzai@71-217-108-78.tukw.qwest.net, Quit: Leaving) | |
| 01:29 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 260 seconds) | |
| 01:42 | fuzai has joined IRC (fuzai!~fuzai@71-217-108-78.tukw.qwest.net) | |
| 01:42 | andygraybeal has left IRC (andygraybeal!~andy@h193.205.130.174.dynamic.ip.windstream.net, Ping timeout: 265 seconds) | |
| 01:43 | adrianorg has left IRC (adrianorg!~adrianorg@177.134.63.247, Ping timeout: 246 seconds) | |
| 01:44 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 02:08 | F-GT has left IRC (F-GT!~phantom@ppp59-167-136-109.static.internode.on.net, Ping timeout: 246 seconds) | |
| 02:14 | dgeary2 has left IRC (dgeary2!~david@2001:388:e001:1e00:914:40fb:752:d15, Ping timeout: 250 seconds) | |
| 02:17 | PhoenixSTF has joined IRC (PhoenixSTF!~rudi@78.29.132.86) | |
| 03:19 | PhoenixSTF has left IRC (PhoenixSTF!~rudi@78.29.132.86, Quit: Leaving) | |
| 03:44 | lostfile has joined IRC (lostfile!~Thunderbi@113.161.68.105) | |
| 04:35 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 252 seconds) | |
| 04:53 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 04:55 | sha has joined IRC (sha!~sha@e177167052.adsl.alicedsl.de) | |
| 04:59 | sha_ has left IRC (sha_!~sha@e177160138.adsl.alicedsl.de, Ping timeout: 255 seconds) | |
| 05:02 | fuzai has left IRC (fuzai!~fuzai@71-217-108-78.tukw.qwest.net, Remote host closed the connection) | |
| 05:33 | telex has left IRC (telex!~telex@freeshell.de, Ping timeout: 265 seconds) | |
| 05:35 | telex has joined IRC (telex!~telex@freeshell.de) | |
| 05:36 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 260 seconds) | |
| 05:52 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 06:31 | F-GT has joined IRC (F-GT!~phantom@ppp59-167-136-109.static.internode.on.net) | |
| 06:31 | bauerski has joined IRC (bauerski!~witekb@frodo.psp.opole.pl) | |
| 07:18 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 07:20 | lostfile has left IRC (lostfile!~Thunderbi@113.161.68.105, Quit: lostfile) | |
| 07:22 | staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Ping timeout: 255 seconds) | |
| 07:25 | shogunx has left IRC (shogunx!~shogunx@2001:4978:106:1:2963:b01d:1c25:5499, Ping timeout: 245 seconds) | |
| 07:26 | staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu) | |
| 07:47 | dobber has joined IRC (dobber!~dobber@213.169.45.222) | |
| 08:04 | khildin has joined IRC (khildin!~khildin@ip-80-236-225-253.dsl.scarlet.be) | |
| 08:50 | Gremble has joined IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com) | |
| 09:31 | gvy has joined IRC (gvy!~mike@altlinux/developer/mike) | |
| 09:46 | adrianorg has joined IRC (adrianorg!~adrianorg@187.113.251.126) | |
| 09:57 | bauerski1 has joined IRC (bauerski1!~witekb@frodo.psp.opole.pl) | |
| 10:02 | Gremble has left IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com, *.net *.split) | |
| 10:02 | bauerski has left IRC (bauerski!~witekb@frodo.psp.opole.pl, *.net *.split) | |
| 10:02 | lotharn has left IRC (lotharn!~nick@24.154.55.32, *.net *.split) | |
| 10:10 | lotharn has joined IRC (lotharn!~nick@24.154.55.32) | |
| 10:11 | Gremble has joined IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com) | |
| 10:23 | alexqwesa_ has joined IRC (alexqwesa_!~alex@alexo-veto.broker.freenet6.net) | |
| 10:23 | leio_ has joined IRC (leio_!~leio@gentoo/developer/leio) | |
| 10:25 | jerryc has joined IRC (jerryc!~pppingme@65.28.110.103) | |
| 10:29 | alexqwesa has left IRC (alexqwesa!~alex@alexo-veto.broker.freenet6.net, Ping timeout: 245 seconds) | |
| 10:29 | mgariepy has left IRC (mgariepy!mgariepy@ubuntu/member/mgariepy, Ping timeout: 245 seconds) | |
| 10:29 | pppingme has left IRC (pppingme!~pppingme@unaffiliated/pppingme, Ping timeout: 245 seconds) | |
| 10:29 | leio has left IRC (leio!~leio@gentoo/developer/leio, Ping timeout: 245 seconds) | |
| 10:29 | mgariepy_ has joined IRC (mgariepy_!mgariepy@nat/revolutionlinux/session) | |
| 10:29 | mgariepy_ has joined IRC (mgariepy_!mgariepy@nat/revolutionlinux/x-tslaouyvflqlyadu) | |
| 10:51 | andygraybeal has joined IRC (andygraybeal!~andy@h193.205.130.174.dynamic.ip.windstream.net) | |
| 10:55 | Parker955_Away has left IRC (Parker955_Away!~parker@74.112.203.151, Ping timeout: 245 seconds) | |
| 11:00 | Parker955_Away has joined IRC (Parker955_Away!~parker@74.112.203.151) | |
| 11:46 | markit has joined IRC (markit!~marco@88-149-177-66.v4.ngi.it) | |
| 11:46 | <markit> hi, any nfs4 guru here? in #nfs no one replies
| |
| 11:49 | <gvy> try mounting #nfs with -o nfsvers=3? :)
| |
| 11:49 | gvy has left IRC (gvy!~mike@altlinux/developer/mike, Quit: goto) | |
| 11:56 | <markit> ok, at the moment I'm unable to umount it sigh
| |
| 11:56 | let's reboot the client
| |
| 11:57 | hours and hours wasted in vain, so sad
| |
| 11:58 | wondering if is bette sync or async
| |
| 11:59 | <alkisg> markit: google translate this one to see how we do it here - it works ok: http://alkisg.mysch.gr/steki/index.php?topic=4847.45
| |
| 12:00 | No nfs-guru here though, we just tried things until it works ok
| |
| 12:00 | <markit> alkisg: thanks :) I've read that nfs4 is much better since caches accesses so I was trying it, but chown does not work from client side in the shared home
| |
| 12:01 | I test with nfs3 now, kde people will not complain, I hope ;P
| |
| 12:02 | <alkisg> Right, the problem we found with nfs4 was uid mapping
| |
| 12:13 | <markit> this post explains the difference, and seems that domain name has to be set and equal for NFSv4: https://blogs.oracle.com/tdh/entry/chown_1_and_username_spaces
| |
| 12:14 | [GuS] has joined IRC ([GuS]!~MysT@unaffiliated/gus/x-663402) | |
| 12:17 | alkisg has left IRC (alkisg!~alkisg@plinet.ioa.sch.gr, Quit: Leaving.) | |
| 12:26 | <markit> mmm access denied by server while mounting. This seems not a good day
| |
| 12:27 | ok, found
| |
| 12:31 | work_alkisg has joined IRC (work_alkisg!~alkisg@plinet.ioa.sch.gr) | |
| 12:33 | <markit> ok, with nfs3 creation in the shared home works
| |
| 12:37 | Gremble has left IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com, Quit: I Leave) | |
| 12:53 | Gremble has joined IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com) | |
| 13:03 | Gremble has left IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com, Quit: I Leave) | |
| 13:03 | khildin has left IRC (khildin!~khildin@ip-80-236-225-253.dsl.scarlet.be, Quit: I'm gone, bye bye) | |
| 13:11 | <markit> mmm is fast!
| |
| 13:11 | * markit perplexed | |
| 13:15 | gvy has joined IRC (gvy!~mike@altlinux/developer/mike) | |
| 13:26 | Gremble has joined IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com) | |
| 13:31 | mgariepy_ has left IRC (mgariepy_!mgariepy@nat/revolutionlinux/x-tslaouyvflqlyadu, Quit: Leaving) | |
| 13:31 | mgariepy has joined IRC (mgariepy!mgariepy@ubuntu/member/mgariepy) | |
| 13:40 | hachque has joined IRC (hachque!quassel@2600:3c01::f03c:91ff:fe96:5060) | |
| 14:10 | garymc has joined IRC (garymc!~chatzilla@host81-148-33-213.in-addr.btopenworld.com) | |
| 14:23 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 14:24 | dgeary2 has joined IRC (dgeary2!~david@2001:388:e001:1e00:914:40fb:752:d15) | |
| 14:31 | bauerski1 has left IRC (bauerski1!~witekb@frodo.psp.opole.pl, Quit: Leaving.) | |
| 14:33 | adrianorg has left IRC (adrianorg!~adrianorg@187.113.251.126, Ping timeout: 250 seconds) | |
| 15:03 | Gremble has left IRC (Gremble!~Ben@cpc29-aztw23-2-0-cust144.18-1.cable.virginmedia.com, Quit: I Leave) | |
| 15:04 | daya has joined IRC (daya!~daya@unaffiliated/daya) | |
| 15:05 | daya has joined IRC (daya!~daya@unaffiliated/daya) | |
| 15:06 | shogunx has joined IRC (shogunx!~shogunx@2001:4978:106:1:84f:77f0:2dc:563c) | |
| 15:06 | alkisg has joined IRC (alkisg!~alkisg@ubuntu/member/alkisg) | |
| 15:11 | bobby_C has joined IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at) | |
| 15:24 | staffencasa has left IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu, Read error: Connection reset by peer) | |
| 15:27 | <markit> alkisg: btw, talking with someone in #kubuntu, I got the suggestion of use zram-config, have you evaluated it?
| |
| 15:28 | <alkisg> markit: I think zram is completely useless for ltsp, where we can use nbd swapping
| |
| 15:28 | <markit> is nbd swapping enabled by default? and isn't /tmp in ram currently?
| |
| 15:28 | <alkisg> It's even harmful as it takes some significant ram percentage. It's only useful if someone doesn't want nbd swapping for some reason
| |
| 15:29 | NBD swapping is on only for low-ram clients
| |
| 15:29 | cat /proc/mounts tells you about the mounted file systems, you can check for /tmp there
| |
| 15:29 | <markit> I see. Current local tests with "async" in fstab instead of my previous "sync" show that kde works fine with /home and /var/tmp on nfs... or I'm doing something wrong. Now I try with ltsp let's see
| |
| 15:33 | <Hyperbyte> Hi alkisg! http://www.set-reizen.nl/kerstkaart/ :-)
| |
| 15:34 | <alkisg> Hi Hyperbyte! Merry xmas + new year for you too :)
| |
| 15:34 | And, nice snow!
| |
| 15:34 | <Hyperbyte> :-D
| |
| 15:38 | <||cw> linux kernel nfs sync code is pretty inefficient
| |
| 15:41 | mdsa has left IRC (mdsa!4454054a@gateway/web/freenode/ip.68.84.5.74, Ping timeout: 245 seconds) | |
| 15:45 | <markit> ||cw: with async is there any risk?
| |
| 15:45 | alkisg: probably the main problem was "sync"
| |
| 15:45 | <||cw> sure there is. make sure you trust your UPS
| |
| 15:45 | <markit> now first login is 140MB (from LDM to responsive desktop)
| |
| 15:45 | and second login is 96MB
| |
| 15:45 | that is much better than older 400MB...
| |
| 15:46 | or maybe kde 4.9.4 has improved something in the meantime
| |
| 15:46 | ||cw: server UPS? otherwyse will I have /home damaged, or just loose last modifications?
| |
| 15:47 | <||cw> just open files with pending writes
| |
| 15:48 | similar situation is enabling write cache on a raid card with no backup battery on it
| |
| 15:48 | <markit> ||cw: mmm with raid card you loose the ENTIRE FS
| |
| 15:48 | with async only last modifications on open files, right?
| |
| 15:48 | <||cw> not always, but it's possible
| |
| 15:49 | * markit now tries sshfs again, who can tell... | |
| 15:49 | <||cw> well, I've not seen any issues, even with a kernel panic, but I wasn't using nfs heavily either
| |
| 15:49 | <markit> mmm so bad, /var/tmp is not populated
| |
| 15:50 | oh, is not because I set it that way, ok, let's revert my settings
| |
| 15:55 | dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174) | |
| 16:06 | dobber has left IRC (dobber!~dobber@213.169.45.222, Remote host closed the connection) | |
| 16:15 | bob has joined IRC (bob!6df2e8ec@gateway/web/freenode/ip.109.242.232.236) | |
| 16:15 | <bob> bob
| |
| 16:16 | bob is now known as Guest43983 | |
| 16:16 | <alkisg> Guest43983: hello
| |
| 16:16 | Guest43983: γράψε: /j #ts.sch.gr
| |
| 16:21 | <Guest43983> ela alki
| |
| 16:21 | me to ALT vrika to koubi
| |
| 16:21 | REMOVE sta agglika
| |
| 16:21 | alla pali zitaei password kai ego malon to exo xasei
| |
| 16:21 | <alkisg> Guest43983: επειδή εδώ είναι μόνο αγγλικό κανάλι, γράψε /j #ts.sch.gr
| |
| 16:22 | risca has joined IRC (risca!~risca@c-135-63.eduroam.liu.se) | |
| 16:23 | Guest43983 is now known as bobptz | |
| 16:26 | <markit> alkisg: probably async is the solution to my problems, making kde able to work good enough, see also http://www.immv.es/articulos-web/tips/kde-slow-startup-when-home-is-mounted-over-nfs.html
| |
| 16:26 | I'm setting all the rest as default, rebuilding the image and test again
| |
| 16:26 | <alkisg> markit: async shouldn't make any difference wrt size
| |
| 16:27 | 600 mb vs 100? how can async do that?
| |
| 16:27 | <markit> (last settings were to have the cache under home, now let's try with default /var/tmp and a symlink)
| |
| 16:27 | alkisg: probably caches a lot? I've no idea
| |
| 16:27 | <alkisg> Yes, but sync/async can't tell KDE how much to cache or not
| |
| 16:27 | <markit> "This has been a problem which I hadn't solved through years (since sync behaviour was default) and never had the time to look at it thoroughfully, because I configured GNOME as a default for users. "
| |
| 16:28 | alkisg: I mean, kde DOES a lot of more I/O, but instead of going through lan, nfs client caches locally?
| |
| 16:28 | <alkisg> My problem with KDE wasn't slowness, it was 600 mb on login
| |
| 16:28 | Hmmm yeah that could be right, I don't remember the actual disk size that KDE uses on /var etc
| |
| 16:28 | mikkel has joined IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk) | |
| 16:28 | <markit> alkisg: I'm so happy that I told you now, but better do last test
| |
| 16:30 | risca has left IRC (risca!~risca@c-135-63.eduroam.liu.se, Quit: Lämnar) | |
| 16:37 | bobptz has left IRC (bobptz!6df2e8ec@gateway/web/freenode/ip.109.242.232.236, Quit: Page closed) | |
| 16:50 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Read error: Operation timed out) | |
| 16:54 | <markit> wow, really seem to work now
| |
| 16:55 | alkisg: have you played with other nfs parameter about packet size or something like that?
| |
| 16:55 | ehm, not packet
| |
| 16:55 | nfs4 has things like rsize=8192,wsize=8192
| |
| 16:56 | <alkisg> markit: yes, on atheros NICs we had to specify those... to something like 15xx
| |
| 16:56 | <markit> also use nfs4 should improve performances even more
| |
| 16:56 | <alkisg> On non-atheros NICs, we didn't modify anything
| |
| 16:56 | <markit> alkisg: you are telling about MTU (my mistake)
| |
| 16:56 | what about nfs fine tunings?
| |
| 16:56 | ok, maybe for gnome is "good enough" though
| |
| 16:57 | <alkisg> It's just /home... usually you don't need to pull a lot of MB from there
| |
| 16:57 | E.g. launching an open office document:
| |
| 16:57 | 120 mb read from NBD,
| |
| 16:57 | 1 mb read from /home
| |
| 16:57 | <markit> as far as I remember, with gnome you have 30 MB, I have 96, so 3x worse, there is room for improvement on KDE side, but at least is usable
| |
| 16:57 | <alkisg> ...so it doesn't make much different if it's nfs, sshfs or any other file system
| |
| 16:58 | <markit> mmm yes, you are right, here also probably most of the traffic is in read (apart first login)
| |
| 16:58 | <alkisg> If your students want to encode video from DV cameras, then OK it'll be different
| |
| 16:58 | <markit> lol :)
| |
| 16:59 | are you on holidays soon? just curious, can we expect further ltsp development?
| |
| 16:59 | are you (you ltsp team) evaluating lxde further?
| |
| 16:59 | <alkisg> The "ltsp team" is a loose team of individuals interested in ltsp :)
| |
| 17:00 | And especially these days we lack cooperation
| |
| 17:00 | So every dev that does have free time and the need to do something wrt ltsp, takes the time and does it
| |
| 17:00 | Personally I think I want to take a look at samba 4, for authentication, shared data etc
| |
| 17:01 | <markit> ok, but are you focused in something specific? i.e. implement XYZ to increase ZXY?
| |
| 17:01 | <alkisg> Maybe we can use much of it on schools...
| |
| 17:01 | <markit> good shot, yes
| |
| 17:01 | <alkisg> Then we'll be able to have ltsp servers/ltsp clients/standalone clients have the same central authentication + homes
| |
| 17:01 | And maybe even windows clients too
| |
| 17:02 | Maybe with that, we can retire all our windows servers, even on schools that don't use linux
| |
| 17:03 | For LTSP, I was waiting for the libpam_sshauth implementation and the LDM removal, but it looks like those things are not currently developed so we'll have to wait for a while...
| |
| 17:08 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |
| 17:17 | dgeary2 has left IRC (dgeary2!~david@2001:388:e001:1e00:914:40fb:752:d15, Ping timeout: 245 seconds) | |
| 17:33 | bobby_C has left IRC (bobby_C!~bobby@85-124-22-227.teleworker.xdsl-line.inode.at, Read error: Operation timed out) | |
| 17:40 | jerryc has left IRC (jerryc!~pppingme@65.28.110.103, Quit: Leaving) | |
| 17:40 | jerryc has joined IRC (jerryc!~pppingme@2001:470:1f11:160:223:8bff:fe6d:bad4) | |
| 17:43 | jerryc has left IRC (jerryc!~pppingme@2001:470:1f11:160:223:8bff:fe6d:bad4, Client Quit) | |
| 18:03 | dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Computer has gone to sleep.) | |
| 18:06 | vagrantc has joined IRC (vagrantc!~vagrant@75-150-46-245-Oregon.hfc.comcastbusiness.net) | |
| 18:06 | vagrantc has joined IRC (vagrantc!~vagrant@freegeek/vagrantc) | |
| 18:10 | dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174) | |
| 18:15 | <Hyperbyte> alkisg?
| |
| 18:15 | You around?
| |
| 18:18 | <alkisg> Hyperbyte: pong
| |
| 18:23 | <Hyperbyte> I'm still having the issues with home dirs sticking around after logout
| |
| 18:24 | I'm going to remedy it by moving /home/ to the fileserver, and mount it via NFS
| |
| 18:24 | <alkisg> Hyperbyte: what are you using? thin clients with localapps?
| |
| 18:24 | <Hyperbyte> Is there some debug information you'd like me to gather before I do this?
| |
| 18:24 | No, ltsp-pnp, fat clients.
| |
| 18:24 | This is at Radio Capelle. :)
| |
| 18:25 | <alkisg> Contents of the home dirs, and result of `cat /proc/mounts` could be useful...
| |
| 18:25 | And maybe ps aux | grep $user, after logout
| |
| 18:28 | <Hyperbyte> Okay, well, I can't 100% reproduce this error
| |
| 18:28 | And the home dirs that have stuck around, have been there for quite a while, so they've all been filled by gnome configs, etc.
| |
| 18:29 | I could provide you with output of find . if that's useful, but it'll be a whole lot of stuff
| |
| 18:29 | As for ps aux, no processes are running from the two users in question still
| |
| 18:30 | And cat /proc/mounts still shows a whole bunch of fusefs mounts from both users.
| |
| 18:30 | fuse gvfs, not fuse fs.
| |
| 18:30 | <alkisg> Yup, the output of find would be fine, along with the output of /proc/mounts
| |
| 18:31 | gvfs... any weird file systems in use, like ftp, cameras...?
| |
| 18:32 | I think we should at least force-unmount the home dir after logout, and also tell sshfs to mount the home dir over the existing one even if it's not empty
| |
| 18:32 | That should cause a lot less headaches, until the actual problems are solved
| |
| 18:33 | gvy has left IRC (gvy!~mike@altlinux/developer/mike, Quit: Leaving) | |
| 18:36 | <Hyperbyte> mmm
| |
| 18:36 | Oh wow
| |
| 18:36 | Even better
| |
| 18:36 | I just reproduced it
| |
| 18:37 | So I have a home dir with the exact files that are left after logout
| |
| 18:37 | It's .pulse stuff
| |
| 18:37 | <alkisg> And that's all?
| |
| 18:39 | <Hyperbyte> http://studio.radiocapelle.nl/procmounts.txt
| |
| 18:39 | http://studio.radiocapelle.nl/homedirs.txt
| |
| 18:40 | Notes: janmiddelkoop and sonnyvalies have since the bug appeared logged out and logged back in several times.. Sonny once I believe, me multiple times
| |
| 18:40 | fransvandersman just logged out and the bug has appeared
| |
| 18:40 | Notable things:
| |
| 18:40 | - fransvandersman has .pulse stuff leftover, but nothing stuck in /proc/mounts
| |
| 18:41 | <alkisg> Let's ignore the ones that have logged in since then
| |
| 18:41 | I think in that case, they're using a "local" /home, i.e. a tmpfs one
| |
| 18:41 | <Hyperbyte> More notable info!
| |
| 18:42 | <alkisg> So all their data will be lost on reboot
| |
| 18:42 | <Hyperbyte> They -do- have processes running still
| |
| 18:42 | But - since the users don't exist anymore (duh)00, I didn't notice them.
| |
| 18:42 | <alkisg> What I noticed from KDE as well, was, that sshfs looks like it's starting a bit late
| |
| 18:42 | <Hyperbyte> There's a gnome-keyring thing still running for someone, and pulseaudio stuff for someone else.
| |
| 18:43 | <alkisg> So some daemons get to run and access /home/username before sshfs mounts /home/username from the server
| |
| 18:43 | <Hyperbyte> I could find out which users exactly, but that'll take some finding out
| |
| 18:43 | <alkisg> It's like it's doing a lazy mount or something...
| |
| 18:43 | * markit wakes up | |
| 18:43 | <Hyperbyte> Welcome back markit. ;-)
| |
| 18:43 | alkisg, should I dig into the ps aux to find out what's running exactly and for who?
| |
| 18:44 | <alkisg> Hyperbyte: it would help a bit, if you have time for it
| |
| 18:44 | <markit> was on phone for a lot of time, I've to setup remotely a web server and a ftp server... never setup a ftp server so far
| |
| 18:44 | <Hyperbyte> I can make time for it.
| |
| 18:44 | <alkisg> Also writing all that down in a bug report would help as well
| |
| 18:44 | <Hyperbyte> Can we make the deal that I do figuring out and you do bug report? :)
| |
| 18:44 | <alkisg> Because it's possible that sshfs is to blame, and not ltsp...
| |
| 18:45 | Nope, not at this time... I won't have much time for ltsp for a few days or maybe even weeks more :-/
| |
| 18:45 | <Hyperbyte> Well if you want my opinion, I think LTSP tries to delete user and remove /home/ while some processes are still running for the user
| |
| 18:45 | <alkisg> But that bug is one of the most annoying ltsp bugs currently
| |
| 18:45 | <Hyperbyte> Removing user works, but /home/ stays put with the files that were or are in-use and can't be removed
| |
| 18:46 | <alkisg> Hyperbyte: /home/fransvandersman/.pulse ==> that was created before the sshfs mount, right?
| |
| 18:46 | <Hyperbyte> Does LTSP do a killall -9 for processes that are still running?
| |
| 18:46 | <alkisg> No
| |
| 18:46 | <Hyperbyte> alkisg, no idea.
| |
| 18:46 | <alkisg> I think the problem is that sshfs does a lazy mount
| |
| 18:46 | So some daemons get to write in the tmpfs before the sshfs mount appears
| |
| 18:46 | <Hyperbyte> Maybe LTSP should just do killall -9 <user> and then fork until all processes are gone, before it goes to nuke the user and home dir
| |
| 18:46 | <alkisg> And then, on the next login, the code refuses to do the sshfs mount completely, because files exist there
| |
| 18:47 | No just a forced unmount would be enough
| |
| 18:47 | <Hyperbyte> mhm
| |
| 18:47 | <alkisg> i.e. umount || umount -l
| |
| 18:48 | There are 3-4 bugs left that are very annoying... this is one of them
| |
| 18:48 | Another is the gnome-panel disappearing...
| |
| 18:49 | <Hyperbyte> http://studio.radiocapelle.nl/psaux.txt
| |
| 18:50 | Now you know which files are left after bug appears, and which processes are still running. :)
| |
| 18:50 | I'll see then if I can find some time to file report, if you're too busy
| |
| 18:50 | Meanwhile, I am switching to NFS for /home/
| |
| 18:53 | <dead_inside> v3 or v4?
| |
| 18:54 | <Hyperbyte> No idea.
| |
| 18:54 | Any recommendations?
| |
| 18:54 | <dead_inside> 4
| |
| 18:54 | it supports file locks
| |
| 18:54 | and some programs like chrome/firefox like to lock
| |
| 18:55 | <Hyperbyte> I don't know what default is
| |
| 18:56 | <markit> dead_inside: are you really able to make it work with ltsp?
| |
| 18:56 | <Hyperbyte> Ah. 4 is default.
| |
| 18:56 | <dead_inside> we have been using 4 for about two years now
| |
| 18:56 | <markit> I tried v4 but I had problems when tried to change ownership of files
| |
| 18:56 | <dead_inside> its an ltsp clust
| |
| 18:56 | <Hyperbyte> markit, root squash!
| |
| 18:56 | <dead_inside> cluster*
| |
| 18:56 | <markit> dead_inside: could you suggest me how to setup?
| |
| 18:56 | Hyperbyte: same config for V3 works flasless
| |
| 18:57 | <Hyperbyte> markit, read up on root squashing on NFS. :-)
| |
| 18:57 | <dead_inside> let me see if i can find a good doc markit
| |
| 18:57 | <markit> Hyperbyte: I use no_root_squash
| |
| 18:57 | <Hyperbyte> I believe there have been lots of squashing adjustments/tweaks between v3 and v4.
| |
| 18:57 | markit, you shouldn't use no_root_squash.
| |
| 18:57 | You should let NFS do it's thing.
| |
| 18:58 | <markit> mmm was root_squash and I've found a suggestion to set to no_
| |
| 18:58 | I could do some experiment again then
| |
| 18:58 | <Hyperbyte> It wouldn't surprise me if they removed the no_root_squash option.
| |
| 18:59 | <alkisg> I don't think NFS booting would work then
| |
| 18:59 | For /home, ok, but not for /...
| |
| 19:01 | <Hyperbyte> alkisg, which part of / exactly?
| |
| 19:01 | <alkisg> Hyperbyte: for instance, no suid root programs would work, since they'd be suid ...nobody
| |
| 19:01 | And I think many daemons refuse to access sockets/files/dirs if they're not owned by root
| |
| 19:02 | <markit> dead_inside: what is your export configs, please?
| |
| 19:02 | <Hyperbyte> alkisg, nfs v4 shows file ownership correctly even if user doesn't have access.
| |
| 19:03 | <dead_inside> markit http://pastebin.com/zA2mUkNf
| |
| 19:03 | <alkisg> Hyperbyte: with root_squash, who would be the owner of /etc/shadow?
| |
| 19:04 | <Hyperbyte> alkisg, depends on who the owner of /etc/shadow is?
| |
| 19:05 | With nfsv4, file permissions don't change with root squash enabled/disabled
| |
| 19:05 | <alkisg> -rw-r----- 1 root shadow 2048 Νοέ 21 21:46 /etc/shadow
| |
| 19:05 | <Hyperbyte> Just the access rights.
| |
| 19:05 | <alkisg> Hyperbyte: the access rights? not the owner?
| |
| 19:05 | Wouldn't that file then be owned by nobody:shadow?
| |
| 19:05 | ...and so it would be a security problem, as anyone would be able to read it?
| |
| 19:06 | <Hyperbyte> No
| |
| 19:06 | I am root@nfs client
| |
| 19:06 | I do an ls -al on /data
| |
| 19:06 | There's a folder drwxrwx--- root:root
| |
| 19:06 | I do cd data... permission denied.
| |
| 19:07 | <alkisg> Can you run a suid root program from nfs?
| |
| 19:07 | <Hyperbyte> I didn't boot from NFS
| |
| 19:07 | <alkisg> It doesn't matter... can you?
| |
| 19:07 | <Hyperbyte> I have no idea?
| |
| 19:07 | <alkisg> I mean, try it... cp one to your /data
| |
| 19:08 | <Hyperbyte> Ah
| |
| 19:08 | That's chmod +u right?
| |
| 19:09 | <alkisg> You can copy an existing program from the path... /bin, /sbin, /usr/sbin...
| |
| 19:10 | <Hyperbyte> Can't you just chmod it?
| |
| 19:10 | <alkisg> Sure but suid won't work with scripts
| |
| 19:10 | So you can't try with e.g. a shell script
| |
| 19:10 | <Hyperbyte> Ah
| |
| 19:11 | Well I copied /bin/cp - but I'm not sure if this has suid bit set?
| |
| 19:11 | How can I check this?
| |
| 19:11 | <alkisg> ls -l /usr/bin | grep sr-
| |
| 19:12 | Anyway /me needs to finish some work, so unfortunately will have to check all these nice things later... bb for now!
| |
| 19:12 | <Hyperbyte> I think the suid bit is lost while copying the file
| |
| 19:12 | <alkisg> cp -a
| |
| 19:12 | <Hyperbyte> cp -a ..yeah
| |
| 19:13 | I don't think this works... or it does and SELinux is preventing it
| |
| 19:13 | Or it's disabled... not entirely sure
| |
| 19:13 | Anyway
| |
| 19:16 | dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Computer has gone to sleep.) | |
| 19:30 | dead_inside has joined IRC (dead_inside!~dead_insi@76.75.3.174) | |
| 19:33 | [GuS] has left IRC ([GuS]!~MysT@unaffiliated/gus/x-663402, Quit: Konversation terminated!) | |
| 19:35 | daya has left IRC (daya!~daya@unaffiliated/daya, Quit: Leaving) | |
| 19:36 | komunista has joined IRC (komunista!~slavko@adsl-195-168-234-074.dynamic.nextra.sk) | |
| 19:42 | <markit> dead_inside: thanks a lot
| |
| 19:43 | ehm, at the end of the discussion that I have not understood, is nfs4 ok and is no_root_squash or root_squash ok?
| |
| 19:44 | Hyperbyte: dead_inside uses no_root_squash also
| |
| 19:46 | alkisg has left IRC (alkisg!~alkisg@ubuntu/member/alkisg, Remote host closed the connection) | |
| 19:47 | gvy has joined IRC (gvy!~mike@altlinux/developer/mike) | |
| 19:54 | <Hyperbyte> markit, that doesn't make it less bad practice.
| |
| 19:54 | According to alkisg, it might be required using it to boot clients via NFS instead of NBD
| |
| 19:54 | <markit> Hyperbyte: I've tried with nfs4 and two stand alone pc
| |
| 19:54 | client was able to create files and change attributes
| |
| 19:54 | but not to change ownership
| |
| 19:55 | (I was root on the client side)
| |
| 19:55 | with nfs3 it worked fine
| |
| 19:55 | <Hyperbyte> But then you only need to set it on the chroot export, which is read-only anyways... you should never do no_root_squash on rw filesystems
| |
| 19:55 | <markit> probably here is explained: https://blogs.oracle.com/tdh/entry/chown_1_and_username_spaces
| |
| 19:56 | <Hyperbyte> That's only when you're not using NIS or LDAP
| |
| 19:56 | <markit> I'm not, in fact
| |
| 19:56 | <Hyperbyte> Well then that makes user mapping very difficult yes.
| |
| 19:57 | <markit> I had the same user with the same id on both parts
| |
| 19:57 | in any case, was just to build a test case for KDE devs to troubleshoot KDE slowliness
| |
| 19:58 | so you think that no_root_squash can be removed for /home sharing and I can try nfs4 instead of 3 for nfs?
| |
| 19:58 | <Hyperbyte> I generally always deploy LDAP, for every setup I have where I need to maintain user accounts.
| |
| 19:58 | <markit> I'll do a test later
| |
| 19:58 | <Hyperbyte> markit, do you know what no_root_squash does exactly?
| |
| 19:58 | <markit> Hyperbyte: I never used it and I tend to avoid layers of complexity
| |
| 19:59 | Hyperbyte: I think uses user "root" instead of "nobody" to access on the server side the share
| |
| 19:59 | or something like that
| |
| 19:59 | <Hyperbyte> So far, LDAP has only made things easier for me. :-)
| |
| 19:59 | <markit> Hyperbyte: in ltsp also?
| |
| 19:59 | <Hyperbyte> Yes.
| |
| 19:59 | <markit> I mean, I'm sure ldap is good for solving problems, just wondering if I have such problems to solve
| |
| 19:59 | I've a standalone ltsp server
| |
| 20:00 | <Hyperbyte> markit, with NFS, user permissions on filesystems are the same on client and server. So if I as a user have access to a dir on the server, I also have access to it on the client.
| |
| 20:00 | So, working from that logic, if I'm root on the client with access to everything, I also have access to everything on the server.
| |
| 20:00 | root squashing is an NFS feature that disables that.
| |
| 20:01 | The "nobody" thing you're seeing instead of root is an NFSv3 feature, NFSv4 doesn't do that, it just shows root, but denies permission regardless.
| |
| 20:11 | My LTSP setup is now using NFSv4 for /home :-)
| |
| 20:15 | mnevans has joined IRC (mnevans!81028195@gateway/web/freenode/ip.129.2.129.149) | |
| 20:19 | sep has left IRC (sep!~sep@40.211.jostedal.no, Ping timeout: 276 seconds) | |
| 20:22 | garymc has left IRC (garymc!~chatzilla@host81-148-33-213.in-addr.btopenworld.com, Quit: ChatZilla 0.9.89 [Firefox 17.0.1/20121128204232]) | |
| 20:35 | <dead_inside> sorry markit i just got back from lunch
| |
| 20:40 | <markit> Hyperbyte: KDE needs /var/tmp access too, so better try, but thanks a lot for tips and clarifications
| |
| 20:50 | vagrantc has left IRC (vagrantc!~vagrant@freegeek/vagrantc, Quit: leaving) | |
| 20:55 | adrianorg has joined IRC (adrianorg!~adrianorg@177.156.56.57) | |
| 21:15 | gvy has left IRC (gvy!~mike@altlinux/developer/mike, Quit: sleep then) | |
| 21:18 | mikkel has left IRC (mikkel!~mikkel@80-71-132-15.u.parknet.dk, Quit: Leaving) | |
| 21:57 | _al_ has left IRC (_al_!~antti@188-67-146-28.bb.dnainternet.fi, Ping timeout: 244 seconds) | |
| 21:59 | _al_ has joined IRC (_al_!~antti@87-93-30-181.bb.dnainternet.fi) | |
| 22:09 | NeonLicht has left IRC (NeonLicht!~NeonLicht@darwin.ugr.es, Ping timeout: 244 seconds) | |
| 22:25 | PhoenixSTF has joined IRC (PhoenixSTF!~rudi@78.29.132.86) | |
| 22:35 | staffencasa has joined IRC (staffencasa!~staffenca@8-220.ptpg.oregonstate.edu) | |
| 22:38 | dead_inside has left IRC (dead_inside!~dead_insi@76.75.3.174, Quit: Leaving...) | |
| 23:02 | komunista has left IRC (komunista!~slavko@adsl-195-168-234-074.dynamic.nextra.sk, Quit: Leaving.) | |
| 23:13 | Phantomas has left IRC (Phantomas!~Phantomas@ubuntu/member/phantomas, Ping timeout: 255 seconds) | |
| 23:29 | Phantomas has joined IRC (Phantomas!~Phantomas@ubuntu/member/phantomas) | |