#ceph IRC Log


IRC Log for 2011-07-28

Timestamps are in GMT/BST.

[0:12] * aliguori (~anthony@cpe-70-123-132-139.austin.res.rr.com) has joined #ceph
[0:21] * lx0 (~aoliva@ has joined #ceph
[0:28] * lxo (~aoliva@19NAACRKS.tor-irc.dnsbl.oftc.net) Quit (Ping timeout: 480 seconds)
[0:28] <sagewk> bchrisman: hope that was helpful! let us know if there's anything we can do on our end.
[0:28] <sagewk> fwiw the shared state can easily be stored with librados, or whatever other trivial clustered/reliable store you want.
[0:29] <bchrisman> ahh yeah??? I was thinking librados for that??? doesn't need to be some huge separate filesystem
[0:30] <bchrisman> I've been fighting some fires in here, which is why I haven't been able to get that NFS testing done.. but things are clearing up a bit.
[0:34] <bchrisman> sagewk: also.. thanks for setting that up??? there's a lot of questions running around about future iscsi stuff due to the linux kernel finally explicitly setting a preference for target drivers. :)
[0:39] * gregorg_taf (~Greg@ has joined #ceph
[0:39] * gregorg (~Greg@ Quit (Read error: Connection reset by peer)
[0:41] * rhyry (~robert.hy@ip-64-111-111-107.dreamhost.com) Quit (Read error: Operation timed out)
[1:20] <sagewk> bchrisman: great, np.
[1:27] * pruby (~tim@leibniz.catalyst.net.nz) Quit (Ping timeout: 480 seconds)
[1:28] * pruby (~tim@leibniz.catalyst.net.nz) has joined #ceph
[2:12] * Tv (~Tv|work@ip-64-111-111-107.dreamhost.com) Quit (Ping timeout: 480 seconds)
[2:27] * cmccabe (~cmccabe@ has left #ceph
[2:38] * joshd (~joshd@ip-64-111-111-107.dreamhost.com) Quit (Quit: Leaving.)
[3:38] * bchrisman (~Adium@70-35-37-146.static.wiline.com) Quit (Quit: Leaving.)
[5:04] * yoshi (~yoshi@p4094-ipngn1601marunouchi.tokyo.ocn.ne.jp) has joined #ceph
[5:15] * darkfaded (~floh@ has joined #ceph
[5:22] * darkfader (~floh@ Quit (Ping timeout: 480 seconds)
[6:20] * bchrisman (~Adium@c-98-207-207-62.hsd1.ca.comcast.net) has joined #ceph
[8:23] * yehuda_hm (~yehuda@99-48-179-68.lightspeed.irvnca.sbcglobal.net) has joined #ceph
[8:49] * jbd (~jbd@ks305592.kimsufi.com) has joined #ceph
[9:18] * jbd (~jbd@ks305592.kimsufi.com) has left #ceph
[9:24] * yehuda_hm (~yehuda@99-48-179-68.lightspeed.irvnca.sbcglobal.net) Quit (Ping timeout: 480 seconds)
[13:06] * yoshi (~yoshi@p4094-ipngn1601marunouchi.tokyo.ocn.ne.jp) Quit (Remote host closed the connection)
[13:24] * phil_ (~quassel@chello080109010223.16.14.vie.surfer.at) has joined #ceph
[16:16] * jim (~chatzilla@astound-69-42-16-6.ca.astound.net) Quit (Ping timeout: 480 seconds)
[16:44] * ftp (~ftp@ has joined #ceph
[16:44] <ftp> hello
[16:45] <ftp> how reliable is the metadataserver of ceph. anybody with experience?
[16:46] <ftp> i am evaluating cluster filesystems for ftp backend
[16:47] <ftp> has anybody used ceph in production here?
[16:51] * yoshi (~yoshi@KD027091032046.ppp-bb.dion.ne.jp) has joined #ceph
[17:18] * yehuda_hm (~yehuda@99-48-179-68.lightspeed.irvnca.sbcglobal.net) has joined #ceph
[17:20] * yoshi (~yoshi@KD027091032046.ppp-bb.dion.ne.jp) Quit (Remote host closed the connection)
[17:44] <bchrisman> ftp: ceph is still not intended for production use at this point.
[17:45] * phil_ (~quassel@chello080109010223.16.14.vie.surfer.at) Quit (Remote host closed the connection)
[17:47] * Tv (~Tv|work@ip-64-111-111-107.dreamhost.com) has joined #ceph
[17:53] * greglap (~Adium@ has joined #ceph
[17:54] * bchrisman (~Adium@c-98-207-207-62.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[18:05] * lx0 (~aoliva@ Quit (Read error: No route to host)
[18:06] * lx0 (~aoliva@9YYAAACMC.tor-irc.dnsbl.oftc.net) has joined #ceph
[18:11] * rsharpe (~Adium@70-35-37-146.static.wiline.com) Quit (Quit: Leaving.)
[18:22] * Juul (~Juul@0x55517066.adsl.cybercity.dk) has joined #ceph
[18:24] * rsharpe (~Adium@70-35-37-146.static.wiline.com) has joined #ceph
[18:34] * aliguori (~anthony@cpe-70-123-132-139.austin.res.rr.com) Quit (Quit: Ex-Chat)
[18:35] * joshd (~joshd@ip-64-111-111-107.dreamhost.com) has joined #ceph
[18:40] * greglap (~Adium@ Quit (Quit: Leaving.)
[18:56] * cmccabe (~cmccabe@c-24-23-254-199.hsd1.ca.comcast.net) has joined #ceph
[18:57] * aliguori (~anthony@ has joined #ceph
[19:02] * bchrisman (~Adium@70-35-37-146.static.wiline.com) has joined #ceph
[19:02] * Juul (~Juul@0x55517066.adsl.cybercity.dk) Quit (Quit: Leaving)
[19:16] * bchrisman (~Adium@70-35-37-146.static.wiline.com) Quit (Quit: Leaving.)
[19:17] * ftp (~ftp@ Quit (Quit: leaving)
[19:23] * bchrisman (~Adium@70-35-37-146.static.wiline.com) has joined #ceph
[19:29] <sagewk> bchrisman: skype issues, missed you. anything new?
[19:33] <bchrisman> sagewk: ahh not yet??? real-soon-now??? gonna pull an allnighter if I can get this friggin thing tested this afternoon.. (problems stemming from internal changes to our build environment which actually are making things much better)
[19:33] <bchrisman> err "can't"
[19:34] <sagewk> ha ok, no pressure from this end, just checking in!
[19:42] * phil_ (~quassel@chello080109010223.16.14.vie.surfer.at) has joined #ceph
[20:23] * gregaf (~Adium@ip-64-111-111-107.dreamhost.com) has left #ceph
[20:23] * gregaf (~Adium@ip-64-111-111-107.dreamhost.com) has joined #ceph
[20:33] * votz_ (~votz@pool-72-78-219-212.phlapa.fios.verizon.net) has joined #ceph
[20:33] * votz (~votz@pool-72-78-219-212.phlapa.fios.verizon.net) Quit (Ping timeout: 480 seconds)
[20:55] * phil_ (~quassel@chello080109010223.16.14.vie.surfer.at) Quit (Remote host closed the connection)
[22:25] <bchrisman> did injectargs syntax change? ceph mds injectargs 0 '--debug-mds 1'
[22:25] <bchrisman> that gives me unrecognized command
[22:25] <Tv> wasn't there a "tell" in there
[22:26] <gregaf> yeah, it changed
[22:26] <bchrisman> I'll fix on wiki
[22:26] <bchrisman> hmm...
[22:26] <bchrisman> new syntax documented somewhere?
[22:27] <bchrisman> ceph mds tell 0 injectargs '--debug-mds 1'
[22:27] * Dantman (~dantman@S0106001731dfdb56.vs.shawcable.net) Quit (Remote host closed the connection)
[22:28] <cmccabe> bchrisman: leave off the dashes?
[22:28] <bchrisman> seemed to work.. or at least say okay?
[22:28] <bchrisman> in the very last form.
[22:28] <gregaf> yeah, I think that's right
[22:28] <gregaf> and you can give the monitors commands too, ceph mon tell 0 injectargs '???bla 0'
[22:28] <gregaf> I think
[22:29] <bchrisman> hmm.. this says okay too: "ceph mds tell 0 injectargs '--debug-mws 1'"
[22:29] <cmccabe> it's being parsed the same way argv is
[22:29] <gregaf> it's reporting okay for the tell part ??? if the injectargs breaks that happens in the MDS itself and isn't reported back
[22:30] <gregaf> or maybe it's reported through the central logger? but not through the tell infrastructure
[22:30] <gregaf> never has been
[22:31] <cmccabe> it actually doesn't say much in mds/MDS.cc
[22:31] <cmccabe> or really anything
[22:32] <cmccabe> the problem is that it uses the same code path as parsing argv, and that code path really doesn't log anything
[22:33] <cmccabe> since we decided to do argv parsing in multiple steps, there's no one point that can say that having something in argv is wrong
[22:33] <cmccabe> so if it's misspelled, it will currently be passed over silently.
[22:34] <cmccabe> it could be augmented to log what it's changed. I guess I'll do that.
[22:34] * jim (~chatzilla@astound-69-42-16-6.ca.astound.net) has joined #ceph
[22:39] <bchrisman> okie..
[22:39] * jim (~chatzilla@astound-69-42-16-6.ca.astound.net) Quit ()
[22:39] <Tv> cmccabe: that's quite miserable..
[22:39] <Tv> the not complaining about wrong argv's part
[22:40] <cmccabe> tv: well, all the programs I write give an error message when there is leftover junk in argv
[22:40] <cmccabe> tv: I changed a few to do that over time
[22:40] <cmccabe> tv: but pretty much everything that was here before I joined will silently ignore garbage on the command line, as far as I know
[22:41] * jim (~chatzilla@astound-69-42-16-6.ca.astound.net) has joined #ceph
[22:42] <cmccabe> tv: anyway, printing out what is changing in injectargs should be do-able, and make that command a lot more usable.
[22:42] <bchrisman> that would be nice yeah.
[22:43] <cmccabe> tv: although even then there are serious problems with injectargs as it now exists. Like I pointed out a week or two ago, if your code is not using the config observer framework, it is not threadsafe.
[22:43] <cmccabe> tv: however, I think the main thing people use injectargs for is changing log levels, and that is now threadsafe.
[22:44] * jim (~chatzilla@astound-69-42-16-6.ca.astound.net) Quit ()
[22:46] <sagewk> the main issue with 'tell' anything (including injectargs) is that it just sends a message to the daemon. there's no reply that works it's way back to the user. it's not a request/response, so you will only see arg parsing errors in the daemon log
[22:46] * Dantman (~dantman@S0106001731dfdb56.vs.shawcable.net) has joined #ceph
[22:47] <cmccabe> sagewk: I think we should print out what has changed in the daemon log though. That would at least be helpful for debugging.
[22:48] <sagewk> sure. and/or print out that warning on unparsed stuff in argv
[22:48] <cmccabe> we could do that too
[22:51] * jim (~chatzilla@astound-69-42-16-6.ca.astound.net) has joined #ceph
[23:06] <cmccabe> just FYI, another thing I think should do is remove admin sockets on SIGTERM
[23:06] <cmccabe> since SIGTERM is generally a controlled shutdown
[23:06] <cmccabe> I guess it's not that important, but it might be nicer?
[23:16] <bchrisman> how would I change ms debug log level via injectargs?
[23:16] <bchrisman> send to mon?
[23:16] <cmccabe> it's tell + injectargs
[23:18] <bchrisman> sorry.. missing something there...
[23:18] <bchrisman> are you saying: ceph mds tell + injectargs '???debug-ms 1'?
[23:18] <gregaf> no???is this for the monitor?
[23:19] <gregaf> ceph mon tell \* injectargs '???debug-ms 1'
[23:19] <gregaf> should do the trik
[23:19] <gregaf> *trick
[23:19] <bchrisman> ahh ok
[23:30] * aliguori (~anthony@ Quit (Ping timeout: 480 seconds)
[23:52] * yoshi (~yoshi@KD027091032046.ppp-bb.dion.ne.jp) has joined #ceph

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