#ceph IRC Log


IRC Log for 2011-08-01

Timestamps are in GMT/BST.

[0:03] * sjust (~sam@ip-64-111-111-107.dreamhost.com) Quit (Ping timeout: 480 seconds)
[0:27] * sjust (~sam@ip-64-111-111-107.dreamhost.com) has joined #ceph
[1:20] * sage (~sage@dsl092-035-022.lax1.dsl.speakeasy.net) Quit (Remote host closed the connection)
[1:21] * sage (~sage@dsl092-035-022.lax1.dsl.speakeasy.net) has joined #ceph
[1:46] * lx0 (~aoliva@09GAAFT8S.tor-irc.dnsbl.oftc.net) Quit (Ping timeout: 481 seconds)
[1:47] * lx0 (~aoliva@1GLAACYCK.tor-irc.dnsbl.oftc.net) has joined #ceph
[2:20] * huangjun (~root@ has joined #ceph
[2:44] * ghaskins (~ghaskins@66-189-113-47.dhcp.oxfr.ma.charter.com) has joined #ceph
[2:54] * aliguori (~anthony@cpe-70-123-132-139.austin.res.rr.com) Quit (Quit: Ex-Chat)
[3:14] * votz (~votz@pool-72-78-219-212.phlapa.fios.verizon.net) has joined #ceph
[3:51] <huangjun> hi,all
[3:52] <huangjun> i really confused, what your read speed? i have 2 osds, and the speed is always 40-50MB/s
[3:53] <huangjun> i need to know, does OSD cached object it reads before? thank you@
[4:21] <huangjun> but ObjectCacher is only worked if we use cfuse or csync, but not kernel client.
[4:55] * phil_ (~quassel@chello080109010223.16.14.vie.surfer.at) Quit (Remote host closed the connection)
[6:51] * bchrisman (~Adium@c-98-207-207-62.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[8:16] * gregorg (~Greg@ Quit (Quit: Quitte)
[8:38] * bchrisman (~Adium@c-98-207-207-62.hsd1.ca.comcast.net) has joined #ceph
[12:23] * hijacker (~hijacker@ Quit (Remote host closed the connection)
[12:27] * hijacker (~hijacker@ has joined #ceph
[13:15] * phil_ (~quassel@chello080109010223.16.14.vie.surfer.at) has joined #ceph
[14:17] * aliguori (~anthony@cpe-70-123-132-139.austin.res.rr.com) has joined #ceph
[14:25] * lx0 (~aoliva@1GLAACYCK.tor-irc.dnsbl.oftc.net) Quit (Ping timeout: 480 seconds)
[14:25] * lx0 (~aoliva@83TAACN97.tor-irc.dnsbl.oftc.net) has joined #ceph
[15:00] <huangjun> good morning
[16:00] * huangjun (~root@ Quit (Remote host closed the connection)
[16:31] * hijacker (~hijacker@ Quit (Quit: Leaving)
[16:32] * hijacker (~hijacker@ has joined #ceph
[17:11] * Tv (~Tv|work@ip-64-111-111-107.dreamhost.com) has joined #ceph
[17:53] * bchrisman (~Adium@c-98-207-207-62.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[17:54] * greglap (~Adium@ has joined #ceph
[18:19] * joshd (~joshd@ip-64-111-111-107.dreamhost.com) has joined #ceph
[18:31] * Juul (~Juul@ has joined #ceph
[18:33] * nolan (~nolan@phong.sigbus.net) Quit (Ping timeout: 480 seconds)
[18:36] * nolan (~nolan@phong.sigbus.net) has joined #ceph
[18:41] * greglap (~Adium@ Quit (Quit: Leaving.)
[18:53] * bchrisman (~Adium@70-35-37-146.static.wiline.com) has joined #ceph
[19:08] * cmccabe (~cmccabe@ has joined #ceph
[19:25] <cmccabe> I guess I'm not sure... what is it that we need to escape in escape_json_attr that we are not currently escaping?
[19:28] * yehudasa (~yehudasa@ip-64-111-111-107.dreamhost.com) Quit (synthon.oftc.net resistance.oftc.net)
[19:28] * [ack]_ (ANONYMOUS@ Quit (synthon.oftc.net resistance.oftc.net)
[19:28] * iggy (~iggy@theiggy.com) Quit (synthon.oftc.net resistance.oftc.net)
[19:28] * u3q (~ben@jupiter.tspigot.net) Quit (synthon.oftc.net resistance.oftc.net)
[19:31] * [ack]_ (ANONYMOUS@ has joined #ceph
[19:31] * u3q (~ben@jupiter.tspigot.net) has joined #ceph
[19:31] * iggy (~iggy@theiggy.com) has joined #ceph
[19:31] * yehudasa (~yehudasa@ip-64-111-111-107.dreamhost.com) has joined #ceph
[19:54] * Juul (~Juul@ Quit (Quit: Leaving)
[19:58] * jmlowe (~Adium@129-79-195-139.dhcp-bl.indiana.edu) has joined #ceph
[20:04] <jmlowe> Does anybody know if the ceph improvements for the 3.0 kernel will be backported to let's say 2.6.38, or more specifically the ubuntu natty narwal kernel?
[20:07] <gregaf> jmlowe: it's unlikely ??? we used to maintain a backports branch but the kernel interfaces went through a lot of churn around that time so we haven't had the manpower to handle backporting
[20:09] <gregaf> I'm not sure how hard 3.0 -> 2.6.38 would be, though, so somebody else might be able to do it, or the appropriate questioning might be inducement enough ??? you'd have to ask Sage when he's not in a meeting :)
[20:09] <jmlowe> so my best choices are build my own from kernel.org or sit tight for oneiric?
[20:11] <jmlowe> build my own 3.0 that is
[20:14] <jmlowe> I guess I could also just shove in the kernel deb's from oneiric prereleases
[20:25] * jojy (~jojyvargh@70-35-37-146.static.wiline.com) has joined #ceph
[20:47] <jojy> is there a doc on the daemon architecture/design. wanted to look at the dentry lookup code
[21:09] <yehudasa> jojy: which daemon?
[21:30] <gregaf> jojy: basically all the documentation we have on the code architecture is in the code, and there's not a lot of it :(
[21:31] <bchrisman> is there doxygen-like output lying around somewhere?
[21:37] <yehudasa> bchrisman: we don't have any
[21:54] <gregaf> hopefully we will have more documentation going forward ??? it is part of our coding style guide now, at least
[22:11] * yehudasa (~yehudasa@ip-64-111-111-107.dreamhost.com) Quit (Ping timeout: 480 seconds)
[22:12] * yehudasa (~yehudasa@ip-64-111-111-107.dreamhost.com) has joined #ceph
[22:17] <jojy> Thanks guys.
[22:24] <jojy> whats the event loop method in mds?
[22:29] <jojy> "dispatch_client_request" looks like it
[22:31] <gregaf> jojy: it's all driven by incoming messages; dispatch_client_request is probably the most important for your purposes but the highest-level one is MDS::ms_dispatch
[22:58] * jmlowe (~Adium@129-79-195-139.dhcp-bl.indiana.edu) Quit (Quit: Leaving.)
[23:02] <sagewk> gregaf, sjust, joshd, cmccabe,tv, yehudasa: quick meeting (~15min) at 4:30
[23:06] <sjust> ok
[23:07] <Tv> sagewk: pushing my overtime for today just over the limit, you'll have to approve that ;)
[23:07] <sagewk> ok!
[23:08] * aliguori (~anthony@cpe-70-123-132-139.austin.res.rr.com) Quit (Read error: Operation timed out)
[23:12] * aliguori (~anthony@cpe-70-123-132-139.austin.res.rr.com) has joined #ceph
[23:23] <jojy> gregaf: although i have mds log level set to 20, i dont see any logs from "dispatch_client_request". Anything quick i shud check
[23:34] <gregaf> jojy: hmm, it should come out with debug mds >= 7
[23:35] <jojy> k
[23:36] <sagewk> yehudasa: cmpxattr is the one right?
[23:37] <yehudasa> hmm
[23:38] <yehudasa> sagewk: there used to be assert_version
[23:38] <sagewk> ah, CEPH_OSD_OP_ASSERT_VER
[23:38] <yehudasa> yep
[23:42] <gregaf> that doesn't actually assert on a failure, does it?
[23:43] <yehudasa> gregaf: no
[23:43] <gregaf> oh good
[23:44] <yehudasa> it asserts the object version, so you can make a compound operation that'll complete only if the object is at a specific version
[23:44] <yehudasa> handy for avoiding races
[23:45] <gregaf> yeah, it makes sense, I was just worried by the use of "assert" in the op title that it was an easy way to do something nasty :)
[23:46] <yehudasa> we can rename it to OSD_OP_CRASH_WHENEVER

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