#ceph IRC Log

Index

IRC Log for 2010-09-30

Timestamps are in GMT/BST.

[0:06] <oksamyt> we'll test this with the release 0.21.3 and hope everything will be fine
[0:06] <oksamyt> thanks again for help!
[0:08] * allsystemsarego (~allsystem@188.27.166.252) Quit (Quit: Leaving)
[0:14] * gregorg_taf (~Greg@78.155.152.6) has joined #ceph
[0:14] * gregorg (~Greg@78.155.152.6) Quit (Read error: Connection reset by peer)
[0:27] * oksamyt (~sana@193.151.59.60) has left #ceph
[0:54] * deksai (~deksai@dsl093-003-018.det1.dsl.speakeasy.net) Quit (Ping timeout: 480 seconds)
[1:08] * Hobbes (~Hobbes@hobbes.user.oftc.net) has joined #ceph
[1:09] * Hobbes (~Hobbes@hobbes.user.oftc.net) Quit (Quit: Leaving)
[1:20] * ani (~Adium@70-35-37-146.static.wiline.com) has joined #ceph
[1:27] * ani (~Adium@70-35-37-146.static.wiline.com) has left #ceph
[1:29] * ani (~Ani@70-35-37-146.static.wiline.com) has joined #ceph
[1:29] * ani (~Ani@70-35-37-146.static.wiline.com) Quit ()
[1:32] * ani (~Ani@70-35-37-146.static.wiline.com) has joined #ceph
[1:33] * ani (~Ani@70-35-37-146.static.wiline.com) has left #ceph
[1:36] * ani (~Ani@70-35-37-146.static.wiline.com) has joined #ceph
[1:50] * ani (~Ani@70-35-37-146.static.wiline.com) has left #ceph
[1:54] * ani (~Ani@70-35-37-146.static.wiline.com) has joined #ceph
[2:04] * ani (~Ani@70-35-37-146.static.wiline.com) Quit (Quit: Leaving.)
[2:04] * ani (~Ani@70-35-37-146.static.wiline.com) has joined #ceph
[2:04] * cmccabe (~cmccabe@dsl081-243-128.sfo1.dsl.speakeasy.net) Quit (Remote host closed the connection)
[2:06] * ani (~Ani@70-35-37-146.static.wiline.com) Quit ()
[3:33] * yehudasa_hm (~yehuda@adsl-69-225-137-176.dsl.irvnca.pacbell.net) has joined #ceph
[4:52] * deksai (~deksai@96-35-100-192.dhcp.bycy.mi.charter.com) has joined #ceph
[5:24] * deksai (~deksai@96-35-100-192.dhcp.bycy.mi.charter.com) Quit (Ping timeout: 480 seconds)
[5:55] * ani (~Ani@netblock-68-183-237-247.dslextreme.com) has joined #ceph
[6:07] * sagelap (~sage@dsl092-035-022.lax1.dsl.speakeasy.net) has joined #ceph
[6:07] * sagelap (~sage@dsl092-035-022.lax1.dsl.speakeasy.net) Quit ()
[6:56] * ani (~Ani@netblock-68-183-237-247.dslextreme.com) Quit (Quit: Leaving.)
[6:58] * f4m8_ is now known as f4m8
[7:50] * yehudasa_hm (~yehuda@adsl-69-225-137-176.dsl.irvnca.pacbell.net) Quit (Ping timeout: 480 seconds)
[8:39] * Yoric (~David@dau94-10-88-189-211-192.fbx.proxad.net) has joined #ceph
[8:44] * allsystemsarego (~allsystem@188.27.166.252) has joined #ceph
[8:58] * Yoric (~David@dau94-10-88-189-211-192.fbx.proxad.net) Quit (Quit: Yoric)
[10:02] * Yoric (~David@213.144.210.93) has joined #ceph
[11:37] * johnl (~johnl@cpc3-brad19-2-0-cust563.barn.cable.virginmedia.com) has joined #ceph
[12:13] * morse (~morse@supercomputing.univpm.it) has joined #ceph
[12:23] * morse (~morse@supercomputing.univpm.it) Quit (Remote host closed the connection)
[12:48] * Yoric (~David@213.144.210.93) Quit (Read error: Connection reset by peer)
[12:48] * Yoric (~David@213.144.210.93) has joined #ceph
[13:22] * Yoric (~David@213.144.210.93) Quit (Read error: Connection reset by peer)
[13:22] * Yoric (~David@213.144.210.93) has joined #ceph
[13:26] * Yoric (~David@213.144.210.93) Quit (Read error: Connection reset by peer)
[14:02] * Yoric (~David@213.144.210.93) has joined #ceph
[14:16] * ani (~Ani@netblock-68-183-237-247.dslextreme.com) has joined #ceph
[15:11] * morse (~morse@supercomputing.univpm.it) has joined #ceph
[15:19] * morse (~morse@supercomputing.univpm.it) Quit (Ping timeout: 480 seconds)
[15:19] * deksai (~deksai@96.35.100.192) has joined #ceph
[15:21] * morse (~morse@supercomputing.univpm.it) has joined #ceph
[15:25] * morse (~morse@supercomputing.univpm.it) Quit (Remote host closed the connection)
[15:26] * morse (~morse@supercomputing.univpm.it) has joined #ceph
[15:28] * morse (~morse@supercomputing.univpm.it) Quit (Read error: Connection reset by peer)
[15:28] * morse (~morse@supercomputing.univpm.it) has joined #ceph
[15:29] * morse (~morse@supercomputing.univpm.it) Quit (Read error: Connection reset by peer)
[15:35] * morse (~morse@supercomputing.univpm.it) has joined #ceph
[15:36] * morse (~morse@supercomputing.univpm.it) Quit (Read error: Connection reset by peer)
[15:38] * morse (~morse@supercomputing.univpm.it) has joined #ceph
[15:39] * morse (~morse@supercomputing.univpm.it) Quit ()
[15:40] * morse (~morse@supercomputing.univpm.it) has joined #ceph
[15:50] * f4m8 is now known as f4m8_
[16:07] * ani (~Ani@netblock-68-183-237-247.dslextreme.com) has left #ceph
[16:35] * allsystemsarego (~allsystem@188.27.166.252) Quit (Quit: Leaving)
[17:44] * Yoric_ (~David@213.144.210.93) has joined #ceph
[17:44] * Yoric (~David@213.144.210.93) Quit (Read error: Connection reset by peer)
[17:44] * Yoric_ is now known as Yoric
[17:50] * deksai (~deksai@96.35.100.192) Quit (Ping timeout: 480 seconds)
[18:12] * greglap (~Adium@166.205.138.121) has joined #ceph
[18:33] * Yoric (~David@213.144.210.93) Quit (Quit: Yoric)
[19:20] * Yoric (~David@dau94-10-88-189-211-192.fbx.proxad.net) has joined #ceph
[19:24] <yehudasa> wido: are you there?
[19:25] * sagelap (~sage@12.248.40.138) has joined #ceph
[19:30] <yehudasa> sagelap: about notify/watch.. didn't we decide to not send any notifications to clients that were disconnected, regardless on when it happened, and make clients send a new watch request with any reconnection?
[19:30] <sagelap> yeah
[19:31] <sagelap> ...so the only place Session*'s need to be linked to ObjectContext's is in the WATCH op code
[19:31] <yehudasa> yeah, and we get rid of anything related to the client once the session dropped
[19:31] <sagelap> yeah
[19:32] <sagelap> btw anything from qemu-devel?
[19:32] <yehudasa> nope
[19:58] * sagelap (~sage@12.248.40.138) Quit (Quit: Leaving.)
[20:06] * greglap (~Adium@166.205.138.121) Quit (Quit: Leaving.)
[20:13] <wido> yehudasa: yes, i'm here
[20:18] <yehudasa> wido: you said that when you try to mount rbd with 'secretfile=' option it crashes?
[20:20] <wido> yes, I'm not sure what I got, but the machine crashed right away
[20:21] <wido> got reboot on panic turned on, so a few minutes later the machine was back
[20:22] <yehudasa> hmm.. what version are you using?
[20:23] <wido> latest, made a clone yesterday
[20:23] <wido> master branch
[20:24] <yehudasa> hmm.. there's no rbd on the master branch, afaik..
[20:24] <yehudasa> or do you mean the ceph.git master branch?
[20:24] * deksai (~deksai@dsl093-003-018.det1.dsl.speakeasy.net) has joined #ceph
[20:24] <wido> no, ceph-clienyt
[20:24] <wido> let me check
[20:25] <wido> Kernel version: 2.6.36-rc5-rbd-20014-g53f0521
[20:25] <yehudasa> I guess that's the rbd branch?
[20:26] <wido> no, unstable branch. I addedd -rbd when making the kernel package
[20:26] <wido> I haven't checked out the rbd branch of ceph-client
[20:26] <yehudasa> the rbd branch is obsolete, don't bother
[20:26] <yehudasa> anyway, I wasn't able to reproduce it here.. is there a way you can get that panic message?
[20:27] <wido> i'll try to add remote syslogging
[20:27] <wido> might catch the error
[20:32] <wido> yehudasa: I see that I already had remote-syslog: http://pastebin.com/62R3FB57
[20:42] <yehudasa> wido: oh, thanks!
[20:42] <yehudasa> I actually managed to reproduce it here now
[20:44] <wido> great. But it's correct that secretfile wasn't supported?
[20:45] <yehudasa> secretfile is not supported anyway.. I think it crashes because of that ipv6 address thing you put in there
[20:46] <yehudasa> currently for the fs, the secretfile parameter is being parsed in the mount.ceph utility
[20:46] <wido> ah, ok. But secretfile would be usefull, like I said, don't want my secret in the bash history
[20:47] <yehudasa> right
[20:48] <yehudasa> you can do something like 'secret=`cat secretfile` for now, would work the same
[20:48] <wido> yes, true :)
[21:02] <yehudasa> wido: I pushed a fix to that crash
[21:04] <wido> ok
[21:04] <wido> what is the easiest way to recompile the modules against the running kernel?
[21:05] <wido> now everything is splitted into the different modules
[21:05] <wido> libceph, ceph and rbd
[21:24] <wido> sagewk: I just read your post in #428. I had a lot of crashed today, but I just saw you had overwritten some binaries on my cluster
[21:24] <wido> but I had overwritten them again this morning after my fresh daily build
[21:25] <wido> Right now one OSD (osd7) keeps crashing, but that might be due to the binaries I've overwritten
[21:36] <yehudasa> wido: you can run multiple 'make SUBDIRS=...' for compiling the different modules
[21:36] <yehudasa> not sure how/if you can do it with a single command
[21:37] <wido> ah, ok, tnx :)
[21:43] <wido> yehudasa: fix confirmed, machine isn't crashing anymore
[21:57] * idletask (~fg@AOrleans-553-1-127-144.w90-21.abo.wanadoo.fr) has joined #ceph
[21:57] <idletask> Good evening
[21:58] <gregaf> hey
[21:58] <idletask> What is the most complete source of documentation to set up a ceph filesystem? The server git repository?
[21:59] <idletask> I have disk space to spare and trust btrfs enough to go the next step
[21:59] <gregaf> you'll want to look at the wiki
[21:59] <gregaf> http://ceph.newdream.net/wiki/Main_Page
[22:00] <gregaf> feel free to ask either here or on the mailing list if stuff isn't clear :)
[22:00] <idletask> Hmm, the link to the wiki wasn't really obvious to see from the main page :p
[22:00] <gregaf> yeah, we should probably make that more obvious
[22:08] <idletask> There are both a fuse-based client and a kernel module to access ceph filesystems - is it for scalability reasons that the kernel module exists, or is it the logical successor to the fuse client?
[22:09] * monrad-51468 (~mmk@domitian.tdx.dk) Quit (Server closed connection)
[22:09] * monrad-51468 (~mmk@domitian.tdx.dk) has joined #ceph
[22:09] <yehudasa> idletask: performance reasons
[22:11] <gregaf> it turns out that all the context switches in FUSE make it sloooow
[22:12] <gregaf> plus the kernel client can be more intelligent about caching and such since it has access access to all the kernel's data about memory pressure, etc
[22:14] <wido> yehudasa: I'm looking into some udev rules to make persistent symlinks to RBD devices
[22:14] <wido> for SCSI devices you have /dev/disk/by-path/XXX
[22:14] <wido> Do the RBD images have a unique ID in them?
[22:15] <wido> right now you can't be sure that /dev/rbd0 always points to the same RBD image, it's just a matter of when it was created
[22:17] <idletask> Don't they have block IDs?
[22:18] <yehudasa> wido: currently we don't export any unique id for them
[22:18] <yehudasa> but there is a unique id for each device
[22:20] <wido> is that ID persistent on all the clients?
[22:20] <yehudasa> yeah
[22:22] <yehudasa> does it need to be a number or a string?
[22:24] <yehudasa> generally, there's the block_name in the rbd header, which is unique per pool
[22:26] <wido> hmm, not sure yet. But what's important, that you can fetch it from /dev/rbd0, so udev can act on it
[22:30] <wido> I'm going afk, i'll see what I can come up with
[22:33] <idletask> The configuration seems to be one of the hardest part for running the thing :p Is multicast considered to go and fetch the nodes' configuration some day?
[23:09] * cmccabe (~cmccabe@dsl081-243-128.sfo1.dsl.speakeasy.net) has joined #ceph
[23:17] * idle_task (~fg@AOrleans-553-1-30-31.w92-152.abo.wanadoo.fr) has joined #ceph
[23:21] * idletask (~fg@AOrleans-553-1-127-144.w90-21.abo.wanadoo.fr) Quit (Ping timeout: 480 seconds)
[23:24] <idle_task> Question about the configuration: is the directory specified by the "mon data" parameter under the [mon] section an absolute path?
[23:26] <gregaf> not unless it's preceded by a /
[23:27] <idle_task> gregaf: and if there is no /, what is it relative to? root's homedir?
[23:27] <gregaf> wherever you run cmon from, IIRC
[23:29] <idle_task> OK, well, I'll continue tomorrow, it's getting late here
[23:29] <idle_task> Have fun!
[23:29] <gregaf> you too :)
[23:29] * idle_task (~fg@AOrleans-553-1-30-31.w92-152.abo.wanadoo.fr) Quit (Quit: .)
[23:46] * Yoric (~David@dau94-10-88-189-211-192.fbx.proxad.net) Quit (Quit: Yoric)

These logs were automatically created by CephLogBot on irc.oftc.net using the Java IRC LogBot.