#ceph IRC Log

Index

IRC Log for 2011-04-25

Timestamps are in GMT/BST.

[16:40] -coulomb.oftc.net- *** Looking up your hostname...
[16:40] -coulomb.oftc.net- *** Checking Ident
[16:40] -coulomb.oftc.net- *** No Ident response
[16:40] -coulomb.oftc.net- *** Found your hostname
[16:40] * CephLogBot (~PircBot@rockbox.widodh.nl) has joined #ceph
[17:16] * greglap1 (~Adium@cpe-76-170-84-245.socal.res.rr.com) Quit (Quit: Leaving.)
[17:19] * neurodrone (~neurodron@cpe-76-180-162-12.buffalo.res.rr.com) has joined #ceph
[17:50] <Yulya> hmmm
[17:51] * greglap (~Adium@198.228.208.139) has joined #ceph
[17:52] <stingray> hmm
[17:53] <Yulya> 2011-04-25 19:50:46.335306 7ffb263fa6f0 ERROR! old-style section name(s) found: mds0, mds1, mds10, mds2, mds3, mds4, mds5, mds.......
[17:54] <Yulya> whats wrong with my config file?
[17:54] <Yulya> http://paste2.org/p/1383869
[17:54] <greglap> Yulya: we switched the naming expectations to be a little stricter
[17:54] <greglap> it wants a period between the daemon type and the id now
[17:54] <greglap> mds.0 instead of mds0
[17:55] <Yulya> oh
[17:55] <Yulya> tnx
[18:06] <Yulya> hm
[18:07] * neurodrone_ (~neurodron@cpe-76-180-162-12.buffalo.res.rr.com) has joined #ceph
[18:07] * neurodrone (~neurodron@cpe-76-180-162-12.buffalo.res.rr.com) Quit (Read error: Connection reset by peer)
[18:07] * neurodrone_ is now known as neurodrone
[18:08] <Yulya> if i upgrade from 0.26 to 0.27 should i create fs again?
[18:10] <greglap> Yulya: it *should* work fine, though it occurs to me you may have some consistency warnings
[18:14] <Yulya> hm
[18:15] <Yulya> ok, i'l try
[18:15] <greglap> we try to maintain backwards compatibility and I don't recall there having been problems with it previously :)
[18:16] <greglap> we added some consistency checks in .27 that turned up some minor issues that you may see warnings on, but that shouldn't do anything but print out some noise in the log
[18:49] * greglap (~Adium@198.228.208.139) Quit (Ping timeout: 480 seconds)
[18:59] * greglap (~Adium@ip-66-33-206-8.dreamhost.com) has joined #ceph
[19:00] * cmccabe (~cmccabe@208.80.64.174) has joined #ceph
[19:01] <cmccabe> yehudasa: are there any object names that aren't valid in rados?
[19:01] <cmccabe> yehudasa: well, never mind, I guess obsync will need to just do whatever rgw would have done
[19:10] * joshd1 (~joshd@ip-66-33-206-8.dreamhost.com) has joined #ceph
[19:14] <bchrisman> cmccabe: what branch are the libceph changes in?
[19:15] <cmccabe> bchrisman: libceph_api
[19:15] <bchrisman> ahh okay.. are those going to be folded into master soon?
[19:15] <cmccabe> bchrisman: yeah
[19:16] <cmccabe> bchrisman: hopefully pretty soon...
[19:19] <bchrisman> cool
[19:19] <cmccabe> bchrisman: actually it looks ready to merge now
[19:19] <bchrisman> yer making my life too easy :)
[19:31] <gregaf> bchrisman: are we supposed to be brining you in today?
[19:32] <bchrisman> yup
[19:32] <bchrisman> we need a conf area.. feh
[19:32] <sagewk> you're not on the scale skype account?
[19:32] <bchrisman> yeah.. our normal place where we have skype installed it taken.. going to get installed on my laptop
[19:32] <bchrisman> but you guys probably need to go ahead
[19:33] <sagewk> k
[19:33] <bchrisman> will see if we can join in time
[19:33] <sagewk> it's dreamhost-aon-conf when you have it
[19:55] <cmccabe> I do have the latest skype installed, btw... the latest for linux, which is Beta Version 2.1.0.81
[19:55] <Tv> cmccabe: we'll just have to send you an upgrade link ;)
[19:55] <cmccabe> heh
[19:56] <cmccabe> I could try running some later version in a VM I guess
[19:56] <cmccabe> assuming that the scale guys have that later version as well
[19:56] <gregaf> that was the SCALE guys this time
[19:56] <gregaf> we got a link as well
[19:57] <cmccabe> we should just have someone set up asterisk...
[19:57] <cmccabe> blah
[19:58] <gregaf> cmccabe: oh bah, I was thinking about it during the meeting and forgot at the end...
[19:58] <gregaf> you're on support watch now :)
[19:58] <cmccabe> ok
[19:58] <Tv> openstack wiki page links go maps.google.fr for the map link
[19:58] <Tv> i was baffled for a while ;)
[19:58] <Tv> s/go/to/
[19:58] <gregaf> guess it's a true international project
[19:59] <gregaf> don't tell me you're Amero-centric! :p
[20:06] * pombreda (~Administr@109.128.137.58) has joined #ceph
[20:12] <Yulya> hm
[20:12] <Yulya> guys, i upgrade my cluster to 0.27
[20:12] <Yulya> and i got the folowing message on mds's http://paste2.org/p/1384098 when i'm trying mount fs from client
[20:14] <cmccabe> yulya: that's bizarre
[20:14] <cmccabe> yulya: len is unsigned?
[20:15] <cmccabe> yulya: oh, never mind, it must be 0
[20:15] <cmccabe> yulya: there were some changes to the mds code recently; that is probably where we have to start debugging this
[20:15] <gregaf> Yulya: please add "debug mds = 20" to your MDS config and try again
[20:16] <Yulya> okay
[20:16] <gregaf> that'll give us more information about the problem
[20:18] <Yulya> hm
[20:18] <Yulya> i dont want to waste your time so...
[20:18] <Yulya> i upgrade without mkcephfs, it may cause this problem?
[20:19] <cmccabe> yulya: what version were you upgrading from?
[20:19] <Yulya> 0.26
[20:19] <cmccabe> yulya: no
[20:19] <Yulya> okay, i'l try debug mds = 20 right now
[20:19] <cmccabe> yulya: you shouldn't need to redo mkcephfs just for that reason
[20:27] <Yulya> http://paste2.org/p/1384133
[20:28] <Yulya> is this enought?
[20:28] <gregaf> yeah
[20:48] <Yulya> hm
[20:49] <Yulya> mkcephfs fixed the problem
[21:22] <gregaf> Yulya: yeah, mkcephfs creates a whole new filesystem so it certainly would :)
[21:22] <gregaf> it looks like it's actually a replay bug around truncation rather than something to do with the upgrade
[21:46] <pombreda> sage, gregaf: howdy :) the playground is down (at least not ssh accessible) since Thursday
[21:47] <pombreda> sry to be a relentless pain ;)
[21:47] <pombreda> and a persistent bastard
[21:48] <gregaf> yeah, we're in the middle of transitioning a lot of its functionality ??? I think Sage was planning to deal with it at some point but you're the only one using it and we have a lot on our plate atm
[21:59] <pombreda> gregaf: no pro, just checking: I am your best fan :)
[22:19] <Tv> note to self: putting "sh -x" in a shell script is very different from "set -x"
[22:20] * Tv facepalms
[22:55] <Tv> FYI: i renamed some ceph-client.git branches to historic/*
[22:55] <Tv> it seems that was enough to make kernel gitbuilder happy, after all
[22:55] <Tv> (as in, i don't need to work around bugs in old code)
[23:07] * greglap1 (~Adium@ip-66-33-206-8.dreamhost.com) has joined #ceph
[23:09] * greglap (~Adium@ip-66-33-206-8.dreamhost.com) Quit (Read error: Connection reset by peer)
[23:16] <sagewk> tv: what is set -x?
[23:16] <Tv> sagewk: echoes commands as they're executed, debug tool
[23:17] <sagewk> oh... same thing as sh -x then!
[23:18] <sagewk> you meant sh -x in the script i take it, not on the #! line
[23:18] <Tv> i mean, if you add a new line "sh -x" into shell script, that starts a subshell that just sits there
[23:18] <Tv> simple typo that boggled me for a while
[23:32] * ghaskins (~ghaskins@66-189-113-47.dhcp.oxfr.ma.charter.com) Quit (Quit: Leaving)
[23:34] * pombreda (~Administr@109.128.137.58) Quit (Quit: Leaving.)
[23:37] <cmccabe> tv: I noticed that the S3 developer's guide states that you can create arbitrary custom metadata key-value pairs in an S3 object at the time of creation
[23:37] <cmccabe> tv: I think perhaps we should add this to the s3 tests
[23:37] <Tv> cmccabe: the headers, yeah
[23:37] <cmccabe> tv: "The metadata is a set of name-value pairs that describe the object. These include default metadata such as the date last modified and standard HTTP metadata such as Content-Type. The developer can also specify custom metadata at the time the Object is stored."
[23:37] <Tv> cmccabe: as i said before, the s3 tests are exactly what Wes had time to do, i only cleaned them up
[23:37] <cmccabe> tv: well, if it doesn't exist now, can we create a bug at least
[23:38] <Tv> cmccabe: basically, add tests as you add features to rgw, or be doomed to fail
[23:38] <cmccabe> tv: I'm not adding a feature to rgw, just trolling through the developer's guide in search of something else
[23:38] <cmccabe> tv: also yehuda claims rgw supports this already
[23:39] <Tv> yeah sounds like it should be added, like many many other things
[23:39] <cmccabe> tv: ok, thanks
[23:39] <gregaf> well everybody's got stuff assigned for this sprint; make a bug for the next sprint and we can prioritize it ;)
[23:42] <sagewk> cmccabe: add the bug.. i think we need to do it this week to get the dho stuff is on solid ground this week.
[23:42] <sagewk> oh i see it
[23:43] <cmccabe> sagewk: added as 1025
[23:43] <cmccabe> going to look at how libboto handles this mess now....
[23:43] <sagewk> k
[23:56] * joshd1 (~joshd@ip-66-33-206-8.dreamhost.com) Quit (Quit: Leaving.)
[23:58] <Tv> configuring company email on my phone.. swype on android insists it's dreamboat.com

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