#ceph IRC Log


IRC Log for 2012-08-07

Timestamps are in GMT/BST.

[0:00] <mgalkiewicz> so the easiest way is to listen on two ips
[0:00] <joshd> could you add new monitors using the new ips, and remove the old monitors when all the clients have switched over?
[0:00] <mgalkiewicz> it is possible but still I am not sure which ip will be reported to clients for osds
[0:01] <joshd> for osds, it's their public address
[0:02] <joshd> as opposed to their cluster address
[0:02] <mgalkiewicz> so the first ip on eth0 interface? how do you choose the address?
[0:03] <joshd> in ceph.conf, 'public addr = ip' for a given osd, or 'public network =' and they'll choose
[0:03] <joshd> clients don't care if osd ips change
[0:04] <mgalkiewicz> could you explain this? what do you suggest?
[0:04] <joshd> since they read it from the osdmap the monitors give them
[0:05] <joshd> if you have static ips for each osd already, just setting their public addr and cluster addr to those ips in ceph.conf and restarting them is the way to go
[0:06] <mgalkiewicz> do you have any description for this on wiki?
[0:06] <mgalkiewicz> I am not sure how to set up public addr and cluster addr in ceph.conf
[0:07] <mgalkiewicz> and how it works
[0:07] <joshd> apparently this isn't in the wiki or docs, but it's definitely been on the mailing list a few times
[0:08] <mgalkiewicz> so this is a sign that it should be added to wiki:)
[0:08] <joshd> indeed :)
[0:09] <joshd> add e.g. 'public addr =' and 'cluster addr =' to the relevant osd section, like [osd.1], in ceph.conf
[0:09] <mgalkiewicz> and public addr and cluster addr represent what?
[0:10] <joshd> frontend and backend network address
[0:11] <joshd> the osd binds to both, but talks to clients on the public addr, and monitors and other osds on the cluster addr
[0:11] <mgalkiewicz> ok but I need that clients uses cluster addr
[0:13] <joshd> if you just change the ip and restart, the monitor will add the new ip to the osdmap and it will get distributed to clients
[0:14] <mgalkiewicz> and the client somehow determines the closest ip?\
[0:14] <mgalkiewicz> like if it is in the same subnet as client it uses the ip?
[0:15] * s[X] (~sX]@ppp59-167-154-113.static.internode.on.net) has joined #ceph
[0:16] <joshd> the client always uses the public address - if you have clients that can't use that new address, there's no way to make it talk to the old one
[0:16] * dspano (~dspano@rrcs-24-103-221-202.nys.biz.rr.com) Quit (Quit: Leaving)
[0:17] <joshd> there may be something you can do with proxying from the old address to the new one (outside of ceph), but I'm not sure how well that would work
[0:19] <mgalkiewicz> ok so your approach is not suitable for ip address change
[0:20] <mgalkiewicz> just like mons
[0:20] <mgalkiewicz> consider some ticket for this
[0:20] * s[X] (~sX]@ppp59-167-154-113.static.internode.on.net) Quit (Remote host closed the connection)
[0:20] <joshd> yeah, not if some clients can't reach the new ip
[0:21] <mgalkiewicz> this is exactly my situation
[0:21] * s[X] (~sX]@ppp59-167-154-113.static.internode.on.net) has joined #ceph
[0:22] <mgalkiewicz> ok thx for help right now I am forcing new clients to use different routes in routing table
[0:22] <mgalkiewicz> but you should consider more flexible way of changing IPs I will probably ask you for more info soon:)
[0:22] <joshd> hehe
[0:23] * mpw (~mpw@chippewa-nat.cray.com) Quit (Quit: Nettalk6 - www.ntalk.de)
[0:23] <joshd> yeah, I'm sure there are other areas the networking could be more flexible too
[0:24] <mgalkiewicz> joshd: thx for help once again I owe you at least six-pack of beers:)
[0:25] <joshd> heh, you're welcome :) a lot of this should be in the docs
[0:27] <dmick> I'm not sure there are many complex network services that handle transitioning IP while still running, are there?
[0:40] <Tv_> FYI status update on sepia: plana & vms are back to normal, burnupi is still on fire
[0:43] <dmick> well many burnupi are ok. just five are hosed (one of which has been down for a while, had a disk replaced and needs reinstall, which is happening)
[0:43] <dmick> nhm is affected; got email out
[0:43] <Tv_> dmick: thanks
[0:43] <Tv_> i just didn't want to claim "burnupi is ok" ;)
[0:43] <dmick> yup
[0:44] <dmick> at this point I wouldn't mind *setting* burnupi on fire, but...
[0:44] <dmick> :)
[0:47] * BManojlovic (~steki@ Quit (Remote host closed the connection)
[0:51] * Leseb (~Leseb@5ED01FAC.cm-7-1a.dynamic.ziggo.nl) Quit (Quit: Leseb)
[0:52] * mgalkiewicz (~mgalkiewi@toya.hederanetworks.net) Quit (Quit: Ex-Chat)
[0:54] <Tv_> dmick: reminds me of the old rule.. always put the UPS on the bottom of the rack, as flames rise up..
[0:54] <dmick> heh
[1:05] <Tv_> gregaf: http://ceph.com/gitbuilder-precise-deb-amd64/output/ref/
[1:12] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) Quit (Ping timeout: 480 seconds)
[1:17] <mikeryan> at my last job the division director strongly implied that it would be easiest to get new switches if "something" happened to the current ones
[1:18] <dmick> "you know, it would be a shame if, I dunno, something...*happened*...to them"
[1:20] * EmilienM (~EmilienM@arc68-4-88-173-120-14.fbx.proxad.net) has left #ceph
[1:24] * Tv_ (~tv@2607:f298:a:607:d976:71b0:669f:be18) Quit (Quit: Tv_)
[1:25] * aliguori (~anthony@cpe-70-123-145-39.austin.res.rr.com) Quit (Remote host closed the connection)
[1:36] * glowell1 (~Adium@ has joined #ceph
[1:38] * glowell (~Adium@2607:f298:a:607:e879:58df:6663:9d8a) Quit (Read error: Connection reset by peer)
[1:46] * Ryan_Lane (~Adium@ has joined #ceph
[1:46] * MarkDude (~MT@c-98-210-253-235.hsd1.ca.comcast.net) Quit (Quit: #BAMF )
[2:00] * tnt_ (~tnt@45.124-67-87.adsl-dyn.isp.belgacom.be) Quit (Ping timeout: 480 seconds)
[2:14] * bchrisman (~Adium@ Quit (Quit: Leaving.)
[2:39] * Cube (~Adium@ Quit (Ping timeout: 480 seconds)
[2:55] * Cube (~Adium@cpe-76-95-223-199.socal.res.rr.com) has joined #ceph
[2:57] * Cube (~Adium@cpe-76-95-223-199.socal.res.rr.com) Quit ()
[3:05] * joshd (~joshd@2607:f298:a:607:221:70ff:fe33:3fe3) Quit (Quit: Leaving.)
[3:32] * themgt (~themgt@cpe-74-78-71-246.maine.res.rr.com) has joined #ceph
[3:33] * themgt (~themgt@cpe-74-78-71-246.maine.res.rr.com) Quit ()
[3:34] * chutzpah (~chutz@ Quit (Quit: Leaving)
[3:39] * adjohn (~adjohn@ Quit (Quit: adjohn)
[3:57] * Cube (~Adium@cpe-76-95-223-199.socal.res.rr.com) has joined #ceph
[4:05] * renzhi (~renzhi@ has joined #ceph
[4:24] * Ryan_Lane (~Adium@ Quit (Quit: Leaving.)
[4:50] * MarkDude (~MT@c-98-210-253-235.hsd1.ca.comcast.net) has joined #ceph
[5:20] * deepsa (~deepsa@ has joined #ceph
[5:45] * adjohn (~adjohn@108-225-130-229.lightspeed.sntcca.sbcglobal.net) has joined #ceph
[5:58] * adjohn (~adjohn@108-225-130-229.lightspeed.sntcca.sbcglobal.net) Quit (Quit: adjohn)
[6:02] * glowell1 (~Adium@ Quit (Quit: Leaving.)
[6:14] * dmick (~dmick@2607:f298:a:607:5f8:127f:6ff7:5407) Quit (Quit: Leaving.)
[6:20] * cattelan (~cattelan@2001:4978:267:0:21c:c0ff:febf:814b) Quit (Ping timeout: 480 seconds)
[6:39] * ssedov (stas@ssh.deglitch.com) has joined #ceph
[6:39] * stass (~stas@ssh.deglitch.com) Quit (Read error: Connection reset by peer)
[6:51] * Ryan_Lane (~Adium@c-67-160-217-184.hsd1.ca.comcast.net) has joined #ceph
[7:08] * morse (~morse@supercomputing.univpm.it) Quit (Remote host closed the connection)
[7:30] * MarkDude (~MT@c-98-210-253-235.hsd1.ca.comcast.net) Quit (Quit: #BAMF )
[7:42] * lxo (~aoliva@lxo.user.oftc.net) Quit (Read error: Connection reset by peer)
[7:56] * lxo (~aoliva@lxo.user.oftc.net) has joined #ceph
[8:06] * deepsa (~deepsa@ Quit (Read error: Connection reset by peer)
[8:06] * deepsa (~deepsa@ has joined #ceph
[8:14] * lyt_lynn (3ad6e9d6@ircip1.mibbit.com) has joined #ceph
[8:22] * adjohn (~adjohn@108-225-130-229.lightspeed.sntcca.sbcglobal.net) has joined #ceph
[8:32] * tnt (~tnt@45.124-67-87.adsl-dyn.isp.belgacom.be) has joined #ceph
[8:38] * deepsa_ (~deepsa@ has joined #ceph
[8:43] * deepsa (~deepsa@ Quit (Ping timeout: 480 seconds)
[8:43] * deepsa_ is now known as deepsa
[9:03] * EmilienM (~EmilienM@arc68-4-88-173-120-14.fbx.proxad.net) has joined #ceph
[9:14] * lyt_lynn (3ad6e9d6@ircip1.mibbit.com) Quit (Quit: http://www.mibbit.com ajax IRC Client)
[9:16] * BManojlovic (~steki@ has joined #ceph
[9:19] * Leseb (~Leseb@2001:980:759b:1:4525:1184:2967:1776) has joined #ceph
[9:19] * s[X] (~sX]@ppp59-167-154-113.static.internode.on.net) Quit (Remote host closed the connection)
[9:22] * Leseb_ (~Leseb@ has joined #ceph
[9:27] * Leseb (~Leseb@2001:980:759b:1:4525:1184:2967:1776) Quit (Ping timeout: 480 seconds)
[9:27] * Leseb_ is now known as Leseb
[9:38] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has joined #ceph
[9:43] * Enigmagic (enigmo@c-24-6-51-229.hsd1.ca.comcast.net) Quit (Read error: Connection reset by peer)
[9:43] <exec> is pgnum is dynamic right now? imagine, that cluster will use 30-40 osds (after creating I see 192 pg/osd), but right now there are only four osd, so pgnum is 192 * 4 = 768.
[9:43] * tnt (~tnt@45.124-67-87.adsl-dyn.isp.belgacom.be) Quit (Ping timeout: 480 seconds)
[9:45] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) has joined #ceph
[9:46] <exec> is it Ok? I didn't found how to change it, but found parameter "mon pg create interval" that means pg autocreation when cluster grows.
[9:53] * verwilst (~verwilst@d5152FEFB.static.telenet.be) has joined #ceph
[9:55] <NaioN> exec: at the moment you can define the number of pg's at creation
[9:56] <NaioN> you can't change it in a running cluster (isn't stable at the moment)
[9:57] <NaioN> ceph osd pool create POOL [pg_num [pgp_num]]
[9:57] <NaioN> with this command you can create a new pool and define the number of pg's
[9:58] <NaioN> the command for adjusting the number of pg's of a pool is in place, but as said isn't stable at the moment
[9:59] * tnt (~tnt@212-166-48-236.win.be) has joined #ceph
[10:02] * adjohn (~adjohn@108-225-130-229.lightspeed.sntcca.sbcglobal.net) Quit (Quit: adjohn)
[10:13] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) Quit (Ping timeout: 480 seconds)
[10:17] <exec> NaioN: thanks, then I'll use separate pools if pg_num will be unacceptable.
[10:55] * morse (~morse@supercomputing.univpm.it) has joined #ceph
[10:56] <NaioN> exec: indeed that's the best option at the moment
[11:04] * Cube (~Adium@cpe-76-95-223-199.socal.res.rr.com) Quit (Quit: Leaving.)
[11:15] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) Quit (Remote host closed the connection)
[11:55] * renzhi (~renzhi@ Quit (Quit: Leaving)
[13:28] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) has joined #ceph
[14:01] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) Quit (Remote host closed the connection)
[14:03] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has joined #ceph
[14:08] * mtk (~mtk@ool-44c35bb4.dyn.optonline.net) has joined #ceph
[14:14] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) has joined #ceph
[14:14] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) Quit (Remote host closed the connection)
[14:16] * shdb (~shdb@80-219-123-230.dclient.hispeed.ch) has joined #ceph
[14:33] * nhorman (~nhorman@hmsreliant.think-freely.org) has joined #ceph
[14:59] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) has joined #ceph
[15:05] * aliguori (~anthony@cpe-70-123-145-39.austin.res.rr.com) has joined #ceph
[15:05] * aliguori (~anthony@cpe-70-123-145-39.austin.res.rr.com) Quit (Remote host closed the connection)
[15:08] * aliguori (~anthony@cpe-70-123-145-39.austin.res.rr.com) has joined #ceph
[15:14] * fiddyspence (~fiddyspen@94-192-234-112.zone6.bethere.co.uk) has joined #ceph
[15:24] <fiddyspence> hi
[15:25] <fiddyspence> i appreciate this is a noddy question, but the docs aren't obvious on this - how do i add nodes to a pre-existing OSD cluster
[15:33] <tnt> http://ceph.com/wiki/OSD_cluster_expansion/contraction
[15:52] <exec> or http://ceph.com/docs/master/ops/manage/grow/
[15:57] * bchrisman (~Adium@c-76-103-130-94.hsd1.ca.comcast.net) has joined #ceph
[16:05] <tnt> Ah yes, better
[16:28] * allsystemsarego (~allsystem@ has joined #ceph
[17:07] * verwilst (~verwilst@d5152FEFB.static.telenet.be) Quit (Quit: Ex-Chat)
[17:13] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) Quit (Remote host closed the connection)
[17:22] * glowell (~Adium@ has joined #ceph
[17:26] * BManojlovic (~steki@ Quit (Quit: Ja odoh a vi sta 'ocete...)
[17:26] * elder (~elder@2607:f298:a:607:3c9c:52cb:843e:71a9) Quit (Quit: Leaving)
[17:26] * Cube (~Adium@ has joined #ceph
[17:27] * loicd1 (~loic@brln-4dbc32b5.pool.mediaWays.net) has joined #ceph
[17:28] * Enigmagic (enigmo@c-24-6-51-229.hsd1.ca.comcast.net) has joined #ceph
[17:29] * mpw (~mpw@chippewa-nat.cray.com) has joined #ceph
[17:31] * lxo (~aoliva@lxo.user.oftc.net) Quit (Ping timeout: 480 seconds)
[17:32] * loicd (~loic@brln-4dbac78d.pool.mediaWays.net) Quit (Ping timeout: 480 seconds)
[17:34] * themgt (~themgt@cpe-74-78-71-246.maine.res.rr.com) has joined #ceph
[17:34] * bchrisman (~Adium@c-76-103-130-94.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[17:35] * Tv_ (~tv@2607:f298:a:607:d976:71b0:669f:be18) has joined #ceph
[17:41] * lxo (~aoliva@lxo.user.oftc.net) has joined #ceph
[17:44] * tnt (~tnt@212-166-48-236.win.be) Quit (Ping timeout: 480 seconds)
[17:56] * elder (~elder@2607:f298:a:607:457f:7231:f46f:bba9) has joined #ceph
[17:58] <elder> I am sitting in Josh's old spot today.
[18:00] * tnt (~tnt@45.124-67-87.adsl-dyn.isp.belgacom.be) has joined #ceph
[18:02] * danieagle (~Daniel@ has joined #ceph
[18:10] * MarkDude (~MT@c-98-210-253-235.hsd1.ca.comcast.net) has joined #ceph
[18:23] * loicd1 is now known as loicd
[18:34] * BManojlovic (~steki@ has joined #ceph
[18:52] * adjohn (~adjohn@ has joined #ceph
[18:56] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) Quit (Ping timeout: 480 seconds)
[18:57] * bchrisman (~Adium@ has joined #ceph
[18:57] * danieagle (~Daniel@ Quit (Quit: Inte+ :-) e Muito Obrigado Por Tudo!!! ^^)
[19:01] * dmick (~dmick@2607:f298:a:607:58ae:bde9:7d72:af37) has joined #ceph
[19:06] <dmick> nhm: ping
[19:07] <dmick> I should check the room list first :)
[19:08] * lofejndif (~lsqavnbok@28IAAGNBI.tor-irc.dnsbl.oftc.net) has joined #ceph
[19:08] * nhorman (~nhorman@hmsreliant.think-freely.org) Quit (Read error: Operation timed out)
[19:09] * chutzpah (~chutz@ has joined #ceph
[19:09] * nhorman (~nhorman@hmsreliant.think-freely.org) has joined #ceph
[19:14] <mikeryan> dmick: bah, who reads lists
[19:14] <mikeryan> just see if tab completion works
[19:18] * Leseb (~Leseb@ Quit (Quit: Leseb)
[19:20] <dmick> yeah, that too
[19:22] * Tobarja (~athompson@cpe-071-075-064-255.carolina.res.rr.com) has joined #ceph
[19:41] * Leseb (~Leseb@5ED01FAC.cm-7-1a.dynamic.ziggo.nl) has joined #ceph
[19:41] * joshd (~joshd@2607:f298:a:607:221:70ff:fe33:3fe3) has joined #ceph
[19:43] * danieagle (~Daniel@ has joined #ceph
[20:08] * Ryan_Lane (~Adium@c-67-160-217-184.hsd1.ca.comcast.net) Quit (Quit: Leaving.)
[20:18] * MarkDude (~MT@c-98-210-253-235.hsd1.ca.comcast.net) Quit (Ping timeout: 480 seconds)
[20:20] <flakrat> FYI, rpmbuild reports two files that are built, but not packaged: /sbin/ceph-disk-activate /sbin/ceph-disk-prepare
[20:21] <flakrat> Can / Should I file a bug report for these?
[20:23] <joshd> yes, those should be packaged
[20:25] <flakrat> I've modified my local copy of the ceph.spec file from the 0.48 tree and verified that they successfully end up in ceph-0.48argonaut-6.el6.x86_64.rpm
[20:25] <flakrat> assuming that's where they belong :-)
[20:27] <dmick> seems rightest to me, flakrat
[20:28] <dmick> they're in the 'ceph' Debian package
[20:31] * danieagle (~Daniel@ Quit (Quit: Inte+ :-) e Muito Obrigado Por Tudo!!! ^^)
[20:47] * MarkDude (~MT@ip-64-134-223-15.public.wayport.net) has joined #ceph
[20:51] * andreask (~andreas@chello062178013131.5.11.vie.surfer.at) has joined #ceph
[21:03] * Ryan_Lane (~Adium@ has joined #ceph
[21:04] * fiddyspence (~fiddyspen@94-192-234-112.zone6.bethere.co.uk) Quit (Quit: Leaving.)
[21:59] * mpw (~mpw@chippewa-nat.cray.com) Quit (Quit: Nettalk6 - www.ntalk.de)
[22:03] * mpw (~mpw@chippewa-nat.cray.com) has joined #ceph
[22:19] * wido (~wido@rockbox.widodh.nl) Quit (Quit: leaving)
[22:20] * wido (~wido@2a00:f10:104:206:9afd:45af:ae52:80) has joined #ceph
[22:22] * nhorman (~nhorman@hmsreliant.think-freely.org) Quit (Quit: Leaving)
[22:27] * wido (~wido@2a00:f10:104:206:9afd:45af:ae52:80) Quit (Quit: leaving)
[22:35] * MapspaM is now known as SpamapS
[22:49] * lxo (~aoliva@lxo.user.oftc.net) Quit (Quit: later)
[22:51] * lxo (~aoliva@lxo.user.oftc.net) has joined #ceph
[22:53] * aliguori (~anthony@cpe-70-123-145-39.austin.res.rr.com) Quit (Remote host closed the connection)
[22:54] * NashTrash (~Adium@mobile-166-147-119-167.mycingular.net) has joined #ceph
[22:54] <NashTrash> Hello Ceph'ers
[22:54] <NashTrash> I am trying to get ceph to auto mount via fstab and I am having some troubles
[22:55] <NashTrash> I have the following entry:
[22:56] <NashTrash> /data ceph name=admin,secret=foo,rw,noexec,nodev,noatime,nodiratime 0 0
[22:56] <NashTrash> It gives me an error in syslog of "device name is missing path (no :/ in"
[22:57] <NashTrash> It works fine with mount -a though
[22:58] * mdxi (~mdxi@74-95-29-182-Atlanta.hfc.comcastbusiness.net) Quit (Quit: leaving)
[22:58] <NashTrash> Also, the documentation at ceph.com/docs/master/cephfs/fstab is wrong
[22:59] * mdxi (~mdxi@74-95-29-182-Atlanta.hfc.comcastbusiness.net) has joined #ceph
[23:00] <iggy> wrong how?
[23:00] * mdxi (~mdxi@74-95-29-182-Atlanta.hfc.comcastbusiness.net) Quit ()
[23:02] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) has joined #ceph
[23:02] <NashTrash> It have noauto in the list of options and ends in 2. Noauto means that mount -a and reboot won't actually mount anything and ending in 2 means run fsck on boot
[23:03] <NashTrash> Minor, but for us copy/pasters it makes it a bit tougher
[23:03] * mdxi (~mdxi@74-95-29-182-Atlanta.hfc.comcastbusiness.net) has joined #ceph
[23:05] <NashTrash> iggy: Have you seen the missing ;/ error before?
[23:05] <iggy> nope
[23:05] <NashTrash> hmm
[23:05] <iggy> did you try just adding another / to see what happens?
[23:06] <NashTrash> mount -a still works and reboot still fails
[23:07] <NashTrash> I just don't see why mount -a would work while the other fails. I thought mount -a was the same a reboot.
[23:07] <iggy> did you try putting netfs?
[23:08] <iggy> that's probably what the person who wrote that was thinking when they put noauto
[23:08] * s[X] (~sX]@ppp59-167-157-96.static.internode.on.net) Quit (Remote host closed the connection)
[23:08] <NashTrash> No, I am not familiar with that. What does that do?
[23:08] * MarkDude (~MT@ip-64-134-223-15.public.wayport.net) Quit (Quit: #BAMF )
[23:09] <iggy> _netdev
[23:09] <iggy> not netfs
[23:10] <NashTrash> Ah, let me give it a try
[23:10] <NashTrash> Ok, mount -a still works. Rebooting...
[23:12] * wido (~wido@2a00:f10:104:206:9afd:45af:ae52:80) has joined #ceph
[23:13] <NashTrash> iggy: Nope. Same error on reboot
[23:14] <NashTrash> Am I able to open a ticket with Inktank?
[23:15] * MarkDude (~MT@ip-64-134-223-15.public.wayport.net) has joined #ceph
[23:16] * JJ (~JJ@ has joined #ceph
[23:18] <dmick> NashTrash: sure
[23:20] <NashTrash> dmick: Do I create a ticket via tracker.newdream.net or somewhere else?
[23:21] <NaioN> NashTrash: is the networking up at the moment of auto-mount?
[23:21] <dmick> http://ceph.com/resources/development/ has a bunch of info, but yes, http://tracker.newdream.net/projects/ceph/issues/new
[23:22] <NaioN> you can provide an option with fstab so it knows it's a fs that depens on networking
[23:22] <NashTrash> NaioN: I added in _netdev which should make sure the net is up. But the error is quite specific about not having a trailing / in the monitor address block
[23:22] <iggy> he tried that
[23:22] <NaioN> oh ok :)
[23:22] <dmick> does it work if you just have ip:port rather than ip:port:/ ?
[23:22] <iggy> did you ever try putting the extra / at the end?
[23:23] <NashTrash> iggy: I did. Mount -a still worked but reboot did not.
[23:23] <NashTrash> Same error message
[23:24] * JohnS50 (~quassel@71-86-129-2.static.stls.mo.charter.com) has joined #ceph
[23:25] <dmick> NashTrash, does it work if you leave off :/ altogether?
[23:25] <NashTrash> dmick: Let me try.
[23:25] <NaioN> NashTrash: is the syntax correct? shouldn't it be: ?
[23:25] <NaioN> the : seperates the port number
[23:26] <NashTrash> Also, here are the syslog entries that contain ceph. Do the first two mean anything...
[23:26] <dmick> NaoiN: also for path, according to mount.ceph(8)
[23:26] <NashTrash> Aug 7 21:16:48 server-2640 kernel: [ 84.560543] libceph: loaded (mon/osd proto 15/24, osdmap 5/6 5/6)
[23:26] <NashTrash> Aug 7 21:16:48 server-2640 kernel: [ 84.569118] ceph: loaded (mds proto 32)
[23:26] <NashTrash> Aug 7 21:16:48 server-2640 kernel: [ 84.571314] ceph: device name is missing path (no :/ in,,
[23:26] <NashTrash> Aug 7 21:16:48 server-2640 kernel: [ 84.668097] ceph: device name is missing path (no :/ in,,
[23:26] <Tv_> NaioN, NashTrash, dmick: is supposed to be the right syntax, but i think we have a bug there, leave the port out
[23:26] <NashTrash> NaioN: I will try leaving the port out.
[23:27] <dmick> I see the code in the filesystem module complaining; something is probably stripping colons wrnog
[23:27] <NaioN> NashTrash: if it's the default port you don't need it anyway...
[23:28] <NashTrash> Nod
[23:28] <NaioN> Tv_: isn't that a bit weird?
[23:29] <NaioN> if you compare it with something like nfs mounting?
[23:29] <dmick> only because nfs doesn't usually allow specifying a port, probably
[23:29] <NashTrash> Tv_: Same error if I leave off the port
[23:29] <NaioN> dmick: true...
[23:29] <Tv_> NashTrash: oh then it's something else.. can you paste the specifics of your last run?
[23:30] <NashTrash> Here is the fstab entry...
[23:30] <NaioN> and ofcourse it's weird that a mount -a works
[23:30] <NashTrash> /data ceph name=vf,secret=foo,rw,noexec,nodev,noatime,nodiratime,_netdev,auto 0 0
[23:30] <NashTrash> Here are the syslog messages...
[23:30] <NashTrash> Aug 7 21:28:54 server-2640 kernel: [ 82.279901] libceph: loaded (mon/osd proto 15/24, osdmap 5/6 5/6)
[23:30] <NashTrash> Aug 7 21:28:54 server-2640 kernel: [ 82.296060] ceph: loaded (mds proto 32)
[23:30] <NashTrash> Aug 7 21:28:54 server-2640 kernel: [ 82.310428] ceph: device name is missing path (no :/ in
[23:30] <NashTrash> Aug 7 21:28:54 server-2640 kernel: [ 82.384112] ceph: device name is missing path (no :/ in
[23:31] <NashTrash> mount -a still works fine
[23:31] <Tv_> dmick: fyi it seems teuthology runs mount.ceph explicitly with ip:port,ip:port,ip:port:/
[23:31] <dmick> oh duh, it's not stripping the :/, it's stripping the /
[23:31] <NashTrash> nod
[23:31] <Tv_> dmick: but i remember this being funky from before, probably due to /etc/fstab parsing
[23:31] <dmick> wonder if :// works around the problem
[23:32] <NashTrash> dmick: tried it. Nope
[23:32] <dmick> right, iggy suggested. hm.
[23:32] <NashTrash> mount -a still works with // though
[23:32] <Tv_> NashTrash: use mount -v to see what it does
[23:33] <NaioN> yeah that's the real strange part
[23:33] <NaioN> why should mount -a parse the fstab different from the auto-mount at boot...
[23:33] <NashTrash> Tv_: sorry for being a noob, but how exactly do you want me to employ mount -v
[23:33] <NashTrash> ?
[23:34] <Tv_> NaioN: oh this is about boot time only? if you run "sudo mount /data" it works?
[23:34] <NashTrash> Tv_: correct. sudo mount /data and mount -a both work. It is only at reboot that it fails.
[23:34] <NaioN> Tv_: NashTrash has the problem that it won't auto-mount at boot as far as i understand
[23:34] <NaioN> so i thought it is a network issue
[23:34] <Tv_> ok that makes this a completely different beast
[23:34] <NaioN> but they used _netdev
[23:35] <NaioN> so something isn't ready at that moment
[23:35] * loicd (~loic@brln-4dbc32b5.pool.mediaWays.net) Quit (Ping timeout: 480 seconds)
[23:35] <dmick> doesn't seem like it's a readiness thing; seems like the string being passed to ceph is wrong in that contet
[23:35] <dmick> *context
[23:35] <Tv_> NaioN: what do you mean used _netdev?
[23:35] <Tv_> NashTrash: what OS is this on?
[23:35] <NaioN> in the fstab as option
[23:36] <Tv_> NaioN: oh right i see it in the later copy paste.. i was looking at the earlier one and was confused
[23:36] <NaioN> so the auto-mounter knows it has to wait on networking before trying to mount the fs
[23:36] <NashTrash> Tv_: A fresh Ubuntu 12.04 LTS Server in an Openstack cloud.
[23:36] <NaioN> but ofcourse something else could not be ready
[23:37] <iggy> I've never had to dig this deep on startup mount problems, does something read that file and then parse that to mount.ceph or does moutn.ceph do the parsing itself (i.e. could this be a nash in the initrd vs bash problem)
[23:37] <NashTrash> NaioN: This would be a very odd error message for something not being ready
[23:37] <NaioN> eg we had a problem with a scsi controller that inits the disks real slow
[23:37] <iggy> or would that have been ruled out with _netdev
[23:37] <NaioN> so we had to wait before we could mount the fs
[23:37] <dmick> yes, 'ready' really does not seem to be the issue
[23:37] <Tv_> iggy: mount execs mount.ceph when fstype=='ceph'
[23:38] <Tv_> i'm worried that bootup is doing something to kludge things
[23:38] <Tv_> reading through the init scripts ;)
[23:38] <Tv_> stuff like mountall(8)
[23:38] <NaioN> dmick: why not?
[23:39] <dmick> because the kernel module is receiving a string with what it believes to be invalid contents, and failing right there
[23:39] <NashTrash> NaioN: I would hope we would get a different error message instead
[23:39] <NaioN> cephfs depends on some modules and networking so everything has to be in place
[23:39] <dmick> it's not trying to access the cluster at the point, it's complaining about the format of thje string it definitely got.
[23:40] <NaioN> yeah but then mount -a shouldn't work either
[23:40] * Ryan_Lane (~Adium@ Quit (Quit: Leaving.)
[23:40] * Ryan_Lane (~Adium@ has joined #ceph
[23:40] <dmick> I'm assuming the automatic mount process is not mount -a
[23:40] <NaioN> so maybe the error isn't correct
[23:40] <dmick> and the parsing of fstab is different
[23:40] <dmick> somehow
[23:40] <dmick> trying to find what the boot process actually does
[23:40] * Ryan_Lane (~Adium@ has left #ceph
[23:41] <Tv_> dmick: no it's mountall(8)
[23:41] <NaioN> yeah as I said I doubt if the parsing is different between mount -a and auto-mount
[23:41] <dmick> NaioN: well we have different doubts :)
[23:41] <NaioN> indeed
[23:41] <NaioN> I still think something isn't ready/initialised
[23:42] <NaioN> it's the most logical
[23:42] <dmick> I understand that you think something isn't ready.
[23:42] <NaioN> and somtimes you get strange errors
[23:42] <dmick> and you may even be right.
[23:42] <dmick> this is quite true.
[23:42] <NashTrash> All, I have to take off. Is there someway I can track this and its possible resolution?
[23:43] <dmick> best way is to file that issue
[23:43] <NashTrash> I registered on newdream tracker but I have not received my confirmation email.
[23:43] <NashTrash> Been 30min or so.
[23:45] <dmick> I'll have to find out how the registration process continues; I've not done a new user setup before; checking
[23:45] <NashTrash> Ok. Thanks. I will open the issue once that is done. Thanks all.
[23:45] * NashTrash (~Adium@mobile-166-147-119-167.mycingular.net) Quit (Quit: Leaving.)
[23:45] * Cube is now known as Concubidated
[23:46] <Tv_> strip_slashes (mnt->device);
[23:46] <Tv_> mountall(8) is worrying me
[23:47] <Tv_> p = str + strlen (str) - 1;
[23:47] <Tv_> while ((p > str) && (*p == '/'))
[23:47] <Tv_> *(p--) = '\0';
[23:47] <Tv_> ok that eats all trailing slashes, right?
[23:48] <Tv_> i think that could be it
[23:55] <JohnS50> I'm having a problem with OSD starting after booting. Can anyone help?

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