Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster, 2014-07-05

| Channels | #gluster index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:05 jonathanpoon joined #gluster
00:57 diegows joined #gluster
01:12 mjsmith2 joined #gluster
01:56 bala joined #gluster
02:55 jag3773 joined #gluster
03:15 sputnik13 joined #gluster
03:22 bennyturns joined #gluster
03:28 sputnik13 joined #gluster
03:40 sputnik13 joined #gluster
04:07 sputnik13 joined #gluster
04:18 RobertLaptop joined #gluster
04:23 sputnik13 joined #gluster
04:45 osiekhan3 joined #gluster
05:20 sputnik13 joined #gluster
05:26 uebera|| joined #gluster
05:39 bala joined #gluster
06:18 jonathanpoon joined #gluster
08:34 sputnik13 joined #gluster
08:47 rjoseph joined #gluster
08:53 hagarth joined #gluster
09:01 ctria joined #gluster
09:05 vimal joined #gluster
10:06 julim joined #gluster
10:11 glusterbot New news from resolvedglusterbugs: [Bug 1088324] glusterfs nfs mount can't handle flock calls. <https://bugzilla.redhat.co​m/show_bug.cgi?id=1088324>
10:17 siel joined #gluster
10:22 xymox joined #gluster
10:39 vimal joined #gluster
10:56 stickyboy joined #gluster
10:57 sijis joined #gluster
10:58 ramteid joined #gluster
11:30 LebedevRI joined #gluster
12:42 diegows joined #gluster
12:48 davinder16 joined #gluster
14:17 DV joined #gluster
14:25 doc|holliday joined #gluster
14:29 doc|holliday I have a cluster of 20+ servers with glusterfs 3.3; want to update it to 3.4.4; I can take one server offline at a time for a short period
14:30 doc|holliday can I mix 3.3 and 3.4.4 in the same cluster? I was thinking disconnect the server, update gluster and rejoin the cluster
14:31 doc|holliday rinse and repeat with all servers
14:32 hagarth doc|holliday: you should be able to do that. Will your clients be accessing data from the volumes during the upgrade?
14:45 doc|holliday hagarth: yes, they will
14:46 doc|holliday btw, this is distributed only cluster -- no redundancy, so some of the data will be inaccessible while the servers are disconnected, which is OK
14:48 JoeJulian doc|holliday: Are the servers also clients?
14:48 doc|holliday JoeJulian: no
14:49 doc|holliday all clients are separate and mount the cluster using fuse
14:50 JoeJulian Then the simple way, "pkill -f gluster", upgrade your packages, then start it again, "service glusterd start" or "service glusterfs-server start" depending on your distro.
14:53 JoeJulian hagarth, can io-stats be disabled? I think 3.4.4 is filling up my log partition with E [iobuf.c:733:iobuf_unref]  (-->/usr/lib64/glusterfs/3.4.4/xlator/d​ebug/io-stats.so(io_stats_readv+0x17d) [0x7f028b5bec3d] (-->/usr/lib64/glusterfs/3.4.4/xlator/p​erformance/md-cache.so(mdc_readv+0x166) [0x7f028b7d36c6] (-->/usr/lib64/glusterfs/3.4.4/xlator/p​erformance/quick-read.so(qr_readv+0x62) [0x7f028b9e0492]))) 0-iobuf: invalid argument: iobuf
14:53 JoeJulian ... or is that quick-read
14:55 doc|holliday JoeJulian: slightly more complicated. I have to re-image the servers, so it will be (a) remove brick, (b) disconnect peer, (c) install new image, (d) probe peer, (e) add brick with existing data <-- btw can I do that?
14:55 JoeJulian Even less simple as you can't remove the peer that has bricks.
14:56 doc|holliday JoeJulian: (a) remove brick
14:56 JoeJulian save /var/lib/glusterd/glusterd.info (at least) or a tar of /var/lib/glusterd....
14:58 doc|holliday JoeJulian: can do that
14:58 JoeJulian You can replace d with restoring that directory and you'll be good to go.
14:59 JoeJulian I missed a completely
14:59 JoeJulian so no removing brick, since you said it's ok for the data to be offline temporarily.
15:00 JoeJulian b,c,replace /var/lib/glusterd, make sure brick is mounted, start glusterd.
15:01 JoeJulian dammit, I need coffee
15:03 JoeJulian backup /var/lib/glusterd, shutdown server, re-image, start server, stop glusterd (if started) restore /var/lib/glusterd, ensure brick is mounted, start glusterd.
15:04 doc|holliday JoeJulian: ok great, that sounds simple enough. thank you sir
15:04 JoeJulian you're welcome
15:13 RicardoSSP joined #gluster
15:25 obelix__ joined #gluster
15:25 glusterbot New news from newglusterbugs: [Bug 1116514] iobuf_unref errors killing logging <https://bugzilla.redhat.co​m/show_bug.cgi?id=1116514>
15:30 zerick joined #gluster
15:30 meghanam joined #gluster
15:40 hagarth JoeJulian: there is no easy way to disable io-stats. Let me take a look at why this is happening.
15:53 hagarth JoeJulian: the log flood is due to quick-read. Sending across a patch.
15:54 hagarth JoeJulian: http://review.gluster.org/8242
15:54 glusterbot Title: Gerrit Code Review (at review.gluster.org)
15:54 mjsmith2 joined #gluster
16:10 bennyturns joined #gluster
16:13 JoeJulian Thanks hagarth
16:31 doekia joined #gluster
16:31 doekia_ joined #gluster
16:34 ricky-ti1 joined #gluster
16:36 social oh hay, I haven't been around for while
16:36 social did anyone looks ince Jun into this http://review.gluster.org/#/c/7223/ ?
16:36 glusterbot Title: Gerrit Code Review (at review.gluster.org)
16:45 delhage joined #gluster
17:06 hagarth social: shyam seems to have some review comments
17:54 mjsmith2 joined #gluster
18:05 andreask joined #gluster
18:37 pureflex joined #gluster
18:37 jag3773 joined #gluster
20:57 glusterbot New news from newglusterbugs: [Bug 1058526] tar keeps reporting "file changed as we read it" on random files <https://bugzilla.redhat.co​m/show_bug.cgi?id=1058526>
21:06 bennyturns joined #gluster
21:40 theron joined #gluster
21:57 diegows joined #gluster
22:31 MacWinner joined #gluster
22:48 MrAbaddon joined #gluster
23:01 fidevo joined #gluster
23:57 sputnik13 joined #gluster

| Channels | #gluster index | Today | | Search | Google Search | Plain-Text | summary