Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster, 2016-05-28

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

All times shown according to UTC.

Time Nick Message
00:01 haomaiwang joined #gluster
00:17 johnmilton joined #gluster
00:38 johnmilton joined #gluster
00:52 MrAbaddon joined #gluster
01:01 EinstCrazy joined #gluster
01:01 haomaiwang joined #gluster
01:08 luizcpg joined #gluster
01:14 alanh joined #gluster
01:15 yosafbridge joined #gluster
01:16 alanh Me again. :-) With my issue of files/directories not showing up via the NFS client after a server crash and restoration of glusterfs/bricks, I found that if I move the files and back they show up. Is there a way I can trigger something in gluster to fix this without having to move the data? It's on a live system and moving and back will cause some issues.
01:16 ashka joined #gluster
01:16 ashka joined #gluster
01:23 ic0n_ joined #gluster
02:01 haomaiwang joined #gluster
02:10 harish joined #gluster
02:10 harish joined #gluster
02:39 harish joined #gluster
02:43 EinstCrazy left #gluster
02:49 noobineer joined #gluster
02:50 natarej joined #gluster
03:01 haomaiwang joined #gluster
04:00 beeradb joined #gluster
04:01 haomaiwang joined #gluster
04:11 Lee1092 joined #gluster
04:46 harish joined #gluster
04:51 haomaiwang joined #gluster
04:58 DV_ joined #gluster
05:28 beeradb joined #gluster
05:29 gnulnx joined #gluster
05:30 gnulnx What would be the difference in creating a four node volume with replica 2, vs a four node volume with replica 4?
05:30 gnulnx I believe the latter would have more of a performance impact as there are 2x more servers to compare files against, write to, etc
05:30 gnulnx but that's as far as my knowledge goes.
05:33 haomaiwang joined #gluster
06:00 om joined #gluster
06:01 haomaiwang joined #gluster
06:05 om2 joined #gluster
06:05 om3 joined #gluster
06:15 om2 joined #gluster
06:30 Siavash___ joined #gluster
06:38 beeradb joined #gluster
06:44 anoopcs hagarth, pong. I just saw your ping. Anything urgent?
06:44 anoopcs I will be available later today.
06:45 beeradb left #gluster
06:59 natarej joined #gluster
07:01 haomaiwang joined #gluster
07:20 Wizek joined #gluster
07:22 nehar joined #gluster
07:47 kovshenin joined #gluster
07:48 CP|AFK joined #gluster
07:48 overclk joined #gluster
07:53 kovshenin joined #gluster
08:01 haomaiwang joined #gluster
08:01 kovshenin joined #gluster
08:09 kovshenin joined #gluster
08:50 haomaiwang joined #gluster
09:03 harish joined #gluster
09:21 haomaiwang joined #gluster
09:31 harish_ joined #gluster
09:45 ahino joined #gluster
09:53 haomaiwang joined #gluster
10:06 ahino joined #gluster
10:24 ahino joined #gluster
10:26 Siavash___ joined #gluster
10:50 haomaiwang joined #gluster
11:01 haomaiwang joined #gluster
11:08 natarej joined #gluster
11:25 atrius joined #gluster
11:30 siel joined #gluster
11:33 Lee1092 joined #gluster
11:34 haomaiwang joined #gluster
11:54 amye joined #gluster
12:12 julim joined #gluster
12:18 Siavash___ joined #gluster
12:36 overclk joined #gluster
12:43 Siavash___ joined #gluster
12:56 overclk joined #gluster
12:57 harish_ joined #gluster
13:04 harish_ joined #gluster
13:20 level7 joined #gluster
13:21 harish_ joined #gluster
13:35 harish_ joined #gluster
13:42 haomaiwang joined #gluster
13:42 nehar joined #gluster
14:01 haomaiwang joined #gluster
14:09 plus joined #gluster
14:09 amye joined #gluster
14:41 hackman joined #gluster
14:56 gnulnx What would be the difference in creating a four node volume with replica 2, vs a four node volume with replica 4?
15:00 Biopandemic joined #gluster
15:01 haomaiwang joined #gluster
15:07 harish_ joined #gluster
15:14 harish_ joined #gluster
15:17 natarej joined #gluster
15:41 mowntan joined #gluster
15:41 amye joined #gluster
16:01 haomaiwang joined #gluster
16:04 jiffin joined #gluster
16:24 Philambdo joined #gluster
16:35 robb_nl joined #gluster
16:39 m0zes joined #gluster
17:01 haomaiwang joined #gluster
17:07 robb_nl joined #gluster
17:15 JesperA joined #gluster
17:25 amye joined #gluster
17:32 JoeJulian gnulnx: With replica 2, the client replicates each file to 2 servers. Replica 4 replicates each file to 4 servers.
17:34 JoeJulian That means that writes will require twice the bandwidth, health checks will require two packets more (miniscule difference), you'll be able to feed the same file to (roughly) twice as many clients simultaneously.
17:35 JoeJulian https://joejulian.name/blog/glusterfs-replication-dos-and-donts/
17:35 glusterbot Title: GlusterFS replication do's and don'ts (at joejulian.name)
17:35 JoeJulian gnulnx: ^
18:01 haomaiwang joined #gluster
18:01 amye joined #gluster
18:20 muneerse2 joined #gluster
18:28 Siavash___ joined #gluster
19:01 haomaiwang joined #gluster
19:19 amye joined #gluster
19:35 level7_ joined #gluster
19:51 F2Knight joined #gluster
20:01 Philambdo joined #gluster
20:01 haomaiwang joined #gluster
20:02 d0nn1e joined #gluster
20:29 deniszh joined #gluster
20:45 manolopm joined #gluster
20:48 manolopm Hi. I've been trying gluster. I followed de Quick Start Guide and its working. I created a brick with a replica in two servers, but I tried a falling one server case, I mean I shutdown one of the servers, create a new file in the mounted fs, and get the second server up. And I ended on dual head case. It's correct? Should I have to activate something to get sync?
20:51 JoeJulian manolopm: You, you did it correctly. What makes you conclude that the file is split-brain?
20:54 manolopm JoeJulian: Because I look into both servers and there are files missing
20:56 manolopm JoeJulian: Now I've tried to see peer status and the server that I don't shutdown said it's disconected from the peer, but from the one I shutdown said connected... I think I'm loosing something
21:01 haomaiwang joined #gluster
21:05 manolopm Ok I found my problem.... iptables. I forgot to clear iptables rules
21:19 chirino joined #gluster
21:20 JoeJulian Excellent
21:21 JoeJulian manolopm: split-brain would be one file that's different on both servers with metadata stating that each had updates for the other. You'll know because "gluster volume heal $vol info" will say, "file in split-brain".
21:23 manolopm JoeJulian: Yes thanks. I have to read more documentation (I'm reading right now) My plan is to create a prototype using gluster as my vmware datastore.
21:34 JoeJulian manolopm: I recommend using ganesha-nfs in an ha configuration (corosync,pacemaker) with gluster for that.
21:38 manolopm JoeJulian: Yes, that's my idea. Thanks
21:45 MikeLupe joined #gluster
21:54 luizcpg joined #gluster
21:56 amye joined #gluster
22:01 haomaiwang joined #gluster
22:14 Wizek joined #gluster
22:24 Wizek_ joined #gluster
23:01 haomaiwang joined #gluster
23:23 Wizek__ joined #gluster
23:35 natarej joined #gluster

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