#ceph IRC Log

Index

IRC Log for 2010-11-23

Timestamps are in GMT/BST.

[1:41] * greglap1 (~Adium@ip-66-33-206-8.dreamhost.com) Quit (Quit: Leaving.)
[1:57] * greglap (~Adium@166.205.138.122) has joined #ceph
[2:57] * greglap (~Adium@166.205.138.122) Quit (Ping timeout: 480 seconds)
[3:03] * cmccabe (~cmccabe@dsl081-243-128.sfo1.dsl.speakeasy.net) Quit (Quit: Leaving.)
[3:06] * greglap (~Adium@cpe-76-90-74-194.socal.res.rr.com) has joined #ceph
[4:46] * greglap (~Adium@cpe-76-90-74-194.socal.res.rr.com) Quit (Quit: Leaving.)
[5:00] * greglap (~Adium@cpe-76-90-74-194.socal.res.rr.com) has joined #ceph
[6:33] * eternaleye_ is now known as eternaleye
[6:42] * f4m8_ is now known as f4m8
[7:28] * andret (~andre@pcandre.nine.ch) Quit (Ping timeout: 480 seconds)
[7:29] * andret (~andre@pcandre.nine.ch) has joined #ceph
[8:07] * ce^thue_thuet^ (~CE_KULIAH@200.199.88.88) has joined #ceph
[8:07] * ce^thue_thuet^ (~CE_KULIAH@200.199.88.88) has left #ceph
[8:10] * ajnelson (~Adium@dhcp-145-95.cruznetsecure.ucsc.edu) has joined #ceph
[8:35] * atg (~atg@please.dont.hacktheinter.net) Quit (Quit: -)
[8:36] * atg (~atg@please.dont.hacktheinter.net) has joined #ceph
[8:49] * Yoric (~David@dau94-10-88-189-211-192.fbx.proxad.net) has joined #ceph
[9:07] * Yoric (~David@dau94-10-88-189-211-192.fbx.proxad.net) Quit (Quit: Yoric)
[9:52] * allsystemsarego (~allsystem@188.26.32.15) has joined #ceph
[10:08] * todinini_ (tuxadero@kudu.in-berlin.de) has joined #ceph
[10:08] * Yoric (~David@213.144.210.93) has joined #ceph
[10:10] * todinini (tuxadero@kudu.in-berlin.de) Quit (Ping timeout: 480 seconds)
[10:37] * ajnelson (~Adium@dhcp-145-95.cruznetsecure.ucsc.edu) Quit (Quit: Leaving.)
[11:04] <jantje_> morning
[11:34] <jantje_> sage: for http://tracker.newdream.net/issues/584, i'm curious what results you get with iozone
[11:35] <jantje_> especially random read/write, both me and wido get MORE random WRITE than random read
[11:37] <jantje_> ./iozone -t 10 -s1G -i 0 -i 2 -i 8
[11:42] * todinini_ (tuxadero@kudu.in-berlin.de) Quit (Remote host closed the connection)
[11:42] * todinini (tuxadero@kudu.in-berlin.de) has joined #ceph
[15:47] * f4m8 is now known as f4m8_
[16:01] * alexxy (~alexxy@79.173.81.171) Quit (Remote host closed the connection)
[16:06] * alexxy (~alexxy@79.173.81.171) has joined #ceph
[16:07] * ajnelson (~Adium@dhcp-145-95.cruznetsecure.ucsc.edu) has joined #ceph
[16:34] * iggy (~iggy@theiggy.com) Quit (Ping timeout: 480 seconds)
[16:35] * nolan (~nolan@phong.sigbus.net) Quit (Ping timeout: 480 seconds)
[16:45] * ajnelson (~Adium@dhcp-145-95.cruznetsecure.ucsc.edu) Quit (Quit: Leaving.)
[17:22] * greglap (~Adium@cpe-76-90-74-194.socal.res.rr.com) Quit (Quit: Leaving.)
[17:24] * iggy_ (~iggy@theiggy.com) has joined #ceph
[17:25] * iggy_ is now known as iggy
[17:28] * Yoric_ (~David@213.144.210.93) has joined #ceph
[17:28] * Yoric (~David@213.144.210.93) Quit (Read error: Connection reset by peer)
[17:28] * Yoric_ is now known as Yoric
[17:51] * greglap (~Adium@166.205.139.203) has joined #ceph
[18:01] * Yoric (~David@213.144.210.93) Quit (Read error: Connection reset by peer)
[18:02] * Yoric (~David@213.144.210.93) has joined #ceph
[18:37] * cmccabe (~cmccabe@dsl081-243-128.sfo1.dsl.speakeasy.net) has joined #ceph
[18:47] * greglap (~Adium@166.205.139.203) Quit (Ping timeout: 480 seconds)
[19:01] * greglap (~Adium@ip-66-33-206-8.dreamhost.com) has joined #ceph
[19:05] * Yoric (~David@213.144.210.93) Quit (Quit: Yoric)
[19:40] * sagewk (~sage@ip-66-33-206-8.dreamhost.com) Quit (Remote host closed the connection)
[19:40] * sagewk (~sage@ip-66-33-206-8.dreamhost.com) has joined #ceph
[19:44] * sagewk (~sage@ip-66-33-206-8.dreamhost.com) Quit (Remote host closed the connection)
[19:55] * sagewk (~sage@ip-66-33-206-8.dreamhost.com) has joined #ceph
[20:06] * Yoric (~David@dau94-10-88-189-211-192.fbx.proxad.net) has joined #ceph
[20:08] <wido> sagewk: I've just read your commit, old fs? This fs was just newly created with the latest unstable
[20:09] <sagewk> created with the exact same version that crashed?
[20:10] <wido> sagewk: yes
[20:11] <wido> "13 KB data, 20680 KB used, 293 GB / 300 GB avail"
[20:11] <wido> there was nothing on it either
[20:13] <sagewk> hmm, can you try to reproduce it?
[20:17] <wido> Yes, sure
[20:17] <wido> i'll rm all the OSD's, and try with the code I have
[20:17] <sagewk> this was on that single machine that you saw this? or the regular cluster?
[20:19] <wido> single machine, small dev box at the office
[20:22] <wido> sagewk: Hmm, weird, I'm sure it was with this code
[20:22] <wido> It could be that I forgot --mkbtrfs with the mkcephfs, could that be it?
[20:23] <wido> that there were some old object around on the OSD
[20:23] * nolan (~nolan@phong.sigbus.net) has joined #ceph
[21:02] <sagewk> hmm... possibly, altho the cosd should clean that up
[21:11] * greglap1 (~Adium@ip-66-33-206-8.dreamhost.com) has joined #ceph
[21:11] * greglap (~Adium@ip-66-33-206-8.dreamhost.com) Quit (Read error: Connection reset by peer)
[21:57] * Meths_ (rift@91.106.208.45) has joined #ceph
[22:01] * Meths (rift@91.106.195.140) Quit (Read error: Operation timed out)
[22:04] * Meths_ is now known as Meths
[22:04] <wido> sagewk: I'm not sure what happend then, pretty weird
[22:04] <iggy> what does the background scrubbing feature do?
[22:04] <wido> btw, could it be that with yesterdays unstable my cluster still doesn't heal completely?
[22:05] <wido> iggy: checks the integrity of your objects compared to other OSD's
[22:05] <iggy> ahh, I was thinking delete/zero/trim/etc. unused objects
[22:07] <wido> Oh, no, it checks if object A is the same on all the OSD's
[22:07] <wido> for example, the object corrupts on a OSD due to some FS error
[22:07] <wido> the scrub will detect that and fix it
[22:11] <wido> sagewk: I'm going afk, you might want to take a look at my regular cluster, it still won't recover (Don't know if you noticed). Tnx again!
[22:12] <gregaf> iggy: it would be hard for RADOS to do something like remove unused objects without support from the user
[22:13] <gregaf> there's an issue in the tracker to support the TRIM command but we haven't implemented it yet
[22:13] <gregaf> scrub itself is still pretty basic, checking for existence and size, but it will get better as time goes on
[22:13] <iggy> if you have repl set to 2, how does it know which one is correct?
[22:14] <gregaf> I don't know what current behavior is on size mis-matches
[22:14] <gregaf> if it exists one place and not the other it's pretty easy to figure out from the PG history whether the object should exist or not
[22:26] <cmccabe> iggy: the master log on the primary will tell it what size the object is
[22:57] <sagewk> wido: still there?
[23:24] <jantje_> what's the issue? 13KB data on a new filesystem ?
[23:27] <jantje_> I think I saw it before, and that it was replaying the journal when first started
[23:27] <jantje_> or something like that
[23:52] * sagelap (~sage@ip-66-33-206-8.dreamhost.com) has joined #ceph
[23:54] * pruby (~tim@leibniz.catalyst.net.nz) Quit (Remote host closed the connection)

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