Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster, 2016-05-24

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

All times shown according to UTC.

Time Nick Message
00:08 mpietersen joined #gluster
00:10 kpease_ joined #gluster
00:19 The_Pugilist joined #gluster
00:19 moss joined #gluster
00:21 lezo joined #gluster
00:34 shaunm joined #gluster
00:38 johnmilton joined #gluster
00:39 mpietersen joined #gluster
00:52 beeradb joined #gluster
00:52 luizcpg joined #gluster
01:06 shyam joined #gluster
01:24 EinstCrazy joined #gluster
01:28 dgandhi joined #gluster
01:30 dgandhi joined #gluster
01:32 dgandhi joined #gluster
01:33 luizcpg joined #gluster
01:34 dgandhi joined #gluster
01:36 dgandhi joined #gluster
01:36 harish joined #gluster
01:38 dgandhi joined #gluster
01:38 Lee1092 joined #gluster
01:40 dgandhi joined #gluster
01:41 jobewan joined #gluster
01:48 ilbot3 joined #gluster
01:48 Topic for #gluster is now Gluster Community - http://gluster.org | Patches - http://review.gluster.org/ | Developers go to #gluster-dev | Channel Logs - https://botbot.me/freenode/gluster/ & http://irclog.perlgeek.de/gluster/
02:00 EinstCrazy joined #gluster
02:10 B21956 joined #gluster
02:15 davpenguin joined #gluster
02:35 pgreg joined #gluster
02:37 pgreg joined #gluster
02:42 haomaiwang joined #gluster
02:50 ninjaryan joined #gluster
03:01 haomaiwang joined #gluster
03:17 ppai joined #gluster
03:26 nehar joined #gluster
03:43 renout_away joined #gluster
03:57 overclk joined #gluster
03:59 itisravi joined #gluster
04:01 haomaiwang joined #gluster
04:10 Wojtek joined #gluster
04:11 atinm joined #gluster
04:14 nishanth joined #gluster
04:20 Gnomethrower joined #gluster
04:23 kotreshhr joined #gluster
04:31 Manikandan joined #gluster
04:39 shortdudey123 joined #gluster
04:42 gem joined #gluster
04:46 aspandey joined #gluster
04:48 DV joined #gluster
05:00 DV__ joined #gluster
05:01 haomaiwang joined #gluster
05:06 Bhaskarakiran joined #gluster
05:08 mowntan joined #gluster
05:11 beeradb_ joined #gluster
05:14 ramky joined #gluster
05:15 tyler274 joined #gluster
05:15 kotreshhr joined #gluster
05:15 malevolent joined #gluster
05:15 kkeithley joined #gluster
05:15 xavih joined #gluster
05:17 raghug joined #gluster
05:17 raghug joined #gluster
05:17 edong23_ joined #gluster
05:18 Ramereth joined #gluster
05:26 Apeksha joined #gluster
05:27 prasanth joined #gluster
05:30 karthik___ joined #gluster
05:31 hgowtham joined #gluster
05:33 jiffin joined #gluster
05:34 nehar joined #gluster
05:39 spalai joined #gluster
05:43 itisravi joined #gluster
05:44 atalur joined #gluster
05:46 ppai joined #gluster
05:48 dgandhi joined #gluster
05:49 The_Ball joined #gluster
05:49 The_Ball What does it mean when volume heal shows: <gfid:71e46e87-aabe-4543-b31a-80c57dee7d44>?
05:49 rafi joined #gluster
05:50 dgandhi joined #gluster
05:50 nehar joined #gluster
05:52 dgandhi joined #gluster
05:53 dgandhi joined #gluster
05:55 aravindavk joined #gluster
05:55 dgandhi joined #gluster
05:57 dgandhi joined #gluster
05:58 itisravi The_Ball: The file corresponding to that gfid needs heal.
05:58 nehar joined #gluster
05:59 The_Ball itisravi, is there a good way to find gfid -> filename mapping?
06:00 dgandhi joined #gluster
06:00 itisravi The_Ball: https://github.com/gluster/glusterfs/bl​ob/master/doc/debugging/gfid-to-path.md
06:00 glusterbot Title: glusterfs/gfid-to-path.md at master · gluster/glusterfs · GitHub (at github.com)
06:01 haomaiwang joined #gluster
06:01 The_Ball Thanks!
06:01 jtux joined #gluster
06:02 skoduri joined #gluster
06:03 ndarshan joined #gluster
06:05 kdhananjay joined #gluster
06:08 Siavash joined #gluster
06:08 Siavash joined #gluster
06:11 jobewan joined #gluster
06:12 Saravanakmr joined #gluster
06:17 cliluw joined #gluster
06:19 tyler274 joined #gluster
06:19 karnan joined #gluster
06:20 valkyr1e joined #gluster
06:22 Pintomatic joined #gluster
06:25 mlhess joined #gluster
06:26 hchiramm joined #gluster
06:28 overclk joined #gluster
06:30 The_Pugilist joined #gluster
06:52 auzty joined #gluster
07:01 haomaiwang joined #gluster
07:03 jri joined #gluster
07:15 armyriad joined #gluster
07:16 [Enrico] joined #gluster
07:17 Saravanakmr joined #gluster
07:20 armyriad joined #gluster
07:21 ivan_rossi joined #gluster
07:23 level7 joined #gluster
07:24 kdhananjay joined #gluster
07:25 Saravanakmr joined #gluster
07:25 Slashman joined #gluster
07:25 ashiq_ joined #gluster
07:26 spalai joined #gluster
07:27 TvL2386 joined #gluster
07:27 level7 joined #gluster
07:32 rastar joined #gluster
07:34 level7 joined #gluster
07:36 fsimonce joined #gluster
07:39 level7 joined #gluster
07:39 anil_ joined #gluster
07:39 skoduri joined #gluster
07:44 level7 joined #gluster
07:44 gowtham_ joined #gluster
07:48 ctria joined #gluster
07:50 RameshN joined #gluster
07:51 level7 joined #gluster
07:51 ahino joined #gluster
07:54 level7 joined #gluster
07:57 karthik___ joined #gluster
07:59 prasanth joined #gluster
08:00 level7 joined #gluster
08:01 haomaiwang joined #gluster
08:08 gowtham_ joined #gluster
08:11 kovshenin joined #gluster
08:18 harish_ joined #gluster
08:24 sdeshpande_ joined #gluster
08:29 arcolife joined #gluster
08:40 EinstCrazy joined #gluster
08:40 TvL2386 joined #gluster
08:41 level7 joined #gluster
08:47 ghenry joined #gluster
08:47 ghenry joined #gluster
08:51 atrius joined #gluster
08:51 JesperA- joined #gluster
08:56 spalai joined #gluster
08:57 ppai joined #gluster
08:57 EinstCrazy joined #gluster
09:01 haomaiwang joined #gluster
09:10 itisravi joined #gluster
09:20 7GHAA6HYK joined #gluster
09:28 deniszh joined #gluster
09:38 Trefex joined #gluster
09:41 shubhendu joined #gluster
09:51 level7 joined #gluster
09:57 itisravi joined #gluster
10:01 haomaiwang joined #gluster
10:05 dgandhi joined #gluster
10:07 dgandhi joined #gluster
10:09 dgandhi joined #gluster
10:10 dgandhi joined #gluster
10:11 JesperA joined #gluster
10:12 dgandhi joined #gluster
10:14 dgandhi joined #gluster
10:15 atalur_ joined #gluster
10:16 dgandhi joined #gluster
10:23 [Enrico] joined #gluster
10:26 rideh joined #gluster
10:32 EinstCrazy joined #gluster
10:48 pgreg_ joined #gluster
11:01 haomaiwang joined #gluster
11:04 DV joined #gluster
11:05 Gnomethrower joined #gluster
11:13 atinm joined #gluster
11:13 ira joined #gluster
11:20 johnmilton joined #gluster
11:21 robb_nl joined #gluster
11:22 hgowtham joined #gluster
11:31 Saravanakmr #REMINDER: Gluster Community Bug Triage meeting at 12:00 UTC ( start ~in 30 minutes)
11:34 karthik___ joined #gluster
11:44 nishanth joined #gluster
11:45 julim joined #gluster
11:48 rafi joined #gluster
12:03 RameshN joined #gluster
12:07 atrius joined #gluster
12:13 nottc joined #gluster
12:16 nehar joined #gluster
12:18 Gnomethrower joined #gluster
12:19 Norky joined #gluster
12:28 fortpedro joined #gluster
12:42 unclemarc joined #gluster
12:46 nishanth joined #gluster
12:46 chirino joined #gluster
12:55 johnmilton joined #gluster
12:58 atinm joined #gluster
13:00 Biopandemic joined #gluster
13:03 nehar joined #gluster
13:06 alvinstarr joined #gluster
13:11 haomaiwang joined #gluster
13:12 Gnomethrower joined #gluster
13:16 karnan joined #gluster
13:16 nishanth joined #gluster
13:19 atrius joined #gluster
13:19 Telsin joined #gluster
13:21 ahino joined #gluster
13:24 skylar joined #gluster
13:28 ic0n_ joined #gluster
13:29 nthomas joined #gluster
13:32 anil_ joined #gluster
13:37 B21956 joined #gluster
13:38 drue joined #gluster
13:38 B21956 joined #gluster
13:38 drue how are people monitoring their gluster clusters in production?
13:40 DV__ joined #gluster
13:43 otaku_coder joined #gluster
13:46 otaku_coder I have a 2-node replicated gluster cluster up and running. I now need to configure the client-side to mount my replicated volume. Documentation for this is somewhat confusing however. Most examples say simply add one of the 2 nodes into your fstab config to mount the volume. I have also found the `backupvolfile-server` parameter which can be passed to the glusterfs-client driver, but then found a changelog w
13:46 otaku_coder hich stated it had been deprecated. My question is this, how do you configure a High-availabiliy client mount?
13:48 Lee1092 joined #gluster
13:50 haomaiwang joined #gluster
14:01 haomaiwang joined #gluster
14:10 P0w3r3d joined #gluster
14:18 post-factum otaku_coder: DNS RR
14:24 atinmu joined #gluster
14:24 atinmu joined #gluster
14:25 dgandhi joined #gluster
14:27 dgandhi joined #gluster
14:29 dgandhi joined #gluster
14:31 aravindavk joined #gluster
14:35 JesperA- joined #gluster
14:45 theeboat joined #gluster
14:45 armyriad joined #gluster
14:47 theeboat Im looking on some advice as to whether gluster will be the correct choice. The content that I am storing is media files ranging from 20-140GB. Over the course of a year we use around 400TB in storage space with the possibility of change. We are looking to buy 60bay chassis with near-line SAS drives, 2x6core Xeon processors & 32GB ram.
14:47 wushudoin joined #gluster
14:47 wushudoin joined #gluster
14:51 rafi1 joined #gluster
15:00 julim joined #gluster
15:01 haomaiwa_ joined #gluster
15:06 mpietersen joined #gluster
15:13 aravindavk joined #gluster
15:17 rtalur joined #gluster
15:22 kpease joined #gluster
15:26 rwheeler joined #gluster
15:39 julim joined #gluster
15:40 ghenry joined #gluster
15:40 fsimonce joined #gluster
15:40 mrEriksson joined #gluster
15:40 yoavz joined #gluster
15:45 atrius joined #gluster
15:45 edong23_ joined #gluster
15:45 jvandewege joined #gluster
15:45 dblack joined #gluster
15:45 XpineX joined #gluster
15:45 javi404 joined #gluster
15:50 skoduri joined #gluster
15:55 timotheus1_ joined #gluster
15:55 kovshenin joined #gluster
16:01 haomaiwang joined #gluster
16:06 kovshenin joined #gluster
16:06 atrius joined #gluster
16:06 edong23_ joined #gluster
16:06 jvandewege joined #gluster
16:06 dblack joined #gluster
16:06 XpineX joined #gluster
16:06 javi404 joined #gluster
16:21 Gnomethrower joined #gluster
16:22 dlambrig_ joined #gluster
16:25 karnan joined #gluster
16:29 armyriad joined #gluster
16:32 jwd joined #gluster
16:32 plarsen joined #gluster
16:40 Siavash joined #gluster
16:40 Siavash joined #gluster
16:46 squizzi_ joined #gluster
16:48 ivan_rossi left #gluster
16:52 mowntan joined #gluster
16:53 mowntan joined #gluster
16:53 mowntan joined #gluster
17:01 haomaiwang joined #gluster
17:05 beeradb_ joined #gluster
17:08 Siavash joined #gluster
17:13 deniszh1 joined #gluster
17:15 luizcpg joined #gluster
17:19 jiffin joined #gluster
17:21 nathwill joined #gluster
17:38 Gnomethrower joined #gluster
17:41 shyam joined #gluster
17:44 tomfite joined #gluster
17:46 tomfite Hey all!
17:57 tomfite Looks like I'm the second person today to ask about production hardware requirements... we're looking to start with ~100 TB cluster containing millions of files of varying sizes. For the storage nodes we're looking at buying at least 8 core with 64GB, 32x4.0TB drives with RAID 6 with a 10GBe card for interconnects. Each box will be set up as one brick.
17:58 tomfite Wondering if anybody knows roughly how many connections a setup like that would be able to deal with.
17:59 JoeJulian Hey, tomfite. Unfortunately theeboat logged off before he could get an answer.
18:00 tomfite Yeah it kinda sounds like we have similar requirements
18:00 JoeJulian I would do that differently with the same hardware (I would add ram though if that can change).
18:01 haomaiwang joined #gluster
18:01 tomfite Would you go to 128 or more?
18:01 JoeJulian 8 x raid0 bricks replica 3 (obviously some multiple of 3 servers)
18:02 hagarth joined #gluster
18:02 JoeJulian I'd buy as much ram as I can get budget for.
18:02 tomfite Can't have too much memory!
18:02 JoeJulian Nobody ever got fired for having stuff in cache.
18:03 tomfite So we're looking at doing replica 3, but I was assuming we'd be replicating across boxes... if we have more than one brick per node can we assume we can lose a box and be OK?
18:03 JoeJulian Yes, across boxes. The reason for multiple bricks is recovery.
18:03 tomfite Oh I see
18:04 tomfite I'm guessing you do RAID 0 for better performance?
18:05 JoeJulian RAID6 increases your MTBF and actually has a longer MTTR.
18:05 JoeJulian Yes, raid 0 just gives you your 10Gb network throughput. If you need 40, adjust as necessary.
18:06 tomfite Gotcha
18:07 JoeJulian So yeah, you lose a drive you lose the whole raid, but you've got 2 replica that are still up on a different machines. Replace the drive, build a new array, let gluster re-replicate.
18:07 ahino joined #gluster
18:07 tomfite OK that makes a lot of sense
18:08 JoeJulian It takes us, with 4TB drives, 10% as long to do it that way as it does to rebuild a 30 disk raid6.
18:08 tomfite Right... we're running RAID 6 currently and the rebuilds take forever
18:08 JoeJulian (well, 28+2)
18:08 tomfite so getting away from that would be great
18:09 JoeJulian Another option, with the new tiering feature, is to not do raid0 at all, but have a hot tier of ssd.
18:10 JoeJulian * I haven't tried tiering yet
18:10 tomfite Interesting, how new is that? I haven't run across it yet.
18:10 JoeJulian http://blog.gluster.org/2016/03​/automated-tiering-in-gluster/
18:10 glusterbot Title: Automated Tiering in Gluster | Gluster Community Website (at blog.gluster.org)
18:10 JoeJulian 3.7
18:11 tomfite Also I was just about to ask you how you make sure that you don't replicate bricks on the same machine but you already answered it https://joejulian.name/blog/how-to-expand-g​lusterfs-replicated-clusters-by-one-server/ :)
18:11 glusterbot Title: How to expand GlusterFS replicated clusters by one server (at joejulian.name)
18:11 tomfite OK cool, I'll dig into that a bit
18:11 JoeJulian That and ,,(brick-order)
18:12 glusterbot I do not know about 'brick-order', but I do know about these similar topics: 'brick order'
18:12 JoeJulian That and ,,(brick order)
18:12 glusterbot Replicas are defined in the order bricks are listed in the volume create command. So gluster volume create myvol replica 2 server1:/data/brick1 server2:/data/brick1 server3:/data/brick1 server4:/data/brick1 will replicate between server1 and server2 and replicate between server3 and server4.
18:12 JoeJulian coffee's slowly beginning to kick in.
18:12 jvandewege_ joined #gluster
18:12 kovsheni_ joined #gluster
18:14 tomfite OK so as long as the bricks are added across machines in volume create (or volume add-brick) then we're good
18:15 tomfite I'll play around with that in my testing environment
18:15 JoeJulian If I had my way there would be three tiers. I wonder if that's possible. I want a frozen tier where the drives are actually powered off.
18:16 edong23 joined #gluster
18:16 javi404 joined #gluster
18:16 JoeJulian When we're running a 19kW rack, it would be nice to be able to shut things down that aren't needed.
18:18 plarsen joined #gluster
18:18 tomfite Efficient!
18:20 anil_ joined #gluster
18:21 tomfite Thanks so much for the help, Joe!
18:23 JoeJulian Any time.
18:23 JoeJulian I'm pretty much always here.
18:30 rafi joined #gluster
18:34 tomfite joined #gluster
18:39 plarsen joined #gluster
18:46 Siavash joined #gluster
18:50 plarsen joined #gluster
18:53 beeradb joined #gluster
18:58 plarsen joined #gluster
18:59 jobewan joined #gluster
19:01 haomaiwang joined #gluster
19:07 P0w3r3d joined #gluster
19:10 deniszh joined #gluster
19:14 shaunm joined #gluster
19:19 P0w3r3d joined #gluster
19:20 ctria joined #gluster
19:25 chirino_m joined #gluster
19:34 ghenry joined #gluster
19:40 beeradb joined #gluster
19:49 Siavash_ joined #gluster
19:57 BitByteNybble110 joined #gluster
20:01 haomaiwang joined #gluster
20:11 DV_ joined #gluster
20:11 luizcpg_ joined #gluster
20:21 edong23 joined #gluster
20:32 kpease joined #gluster
20:43 deniszh joined #gluster
20:45 jri joined #gluster
21:01 haomaiwang joined #gluster
21:02 guhcampos joined #gluster
21:03 guhcampos Is there any other GUI frontend for managing glusterfs besides ovirt? I have some volumes which span to machines which are not part of my ovirt cluster, and ovirt can't see the bricks on those
21:04 guhcampos I'm mostly interested in having a visual overview of the volume's health, but being able to manage the volumes is a plus
21:13 frakt joined #gluster
21:29 Wizek joined #gluster
21:35 johnmilton joined #gluster
21:50 tomfite joined #gluster
21:50 shaunm joined #gluster
21:57 level7 joined #gluster
21:58 plarsen joined #gluster
22:01 haomaiwang joined #gluster
22:07 DV_ joined #gluster
22:16 JoeJulian guhcampos: nope, though the rewrite of glusterd is supposed to have an html api so rolling your own should be pretty easy at that point.
22:16 JoeJulian And you can presently use the cli's xml output to do the same.
22:17 eKKiM_ joined #gluster
22:19 guhcampos JoeJulian I see, thanks. I tried the glusterfs capabilities in ovirt, gave up. It does not really work
22:27 beeradb joined #gluster
22:33 mpietersen joined #gluster
22:42 JoeJulian I've never successfully used ovirt. I tried it out at home for about 4 days before I gave up on it.
23:01 haomaiwang joined #gluster
23:33 fortpedro joined #gluster
23:38 ravana_2 joined #gluster
23:41 ravana_2 Hello all, I'm new to Gluster. I have a 2 node replicated setup. If a node is unavailable, is it normal to see a performance degradation?
23:44 JoeJulian Not usually, no.
23:49 ravana_2 Thanks JoeJulian , I'm testing on AWS and dd showed 94.6 MB/s and after taking a node down writing speed was 644 kB/s.
23:50 JoeJulian I wouldn't use dd for any application you need cluster-wide performance from.
23:51 JoeJulian But yes, that doesn't sound normal.
23:52 JoeJulian I've literally seen increased performance during an outage. It makes sense if you think about it. All the client needs to do during an outage it write to one server and mark the other as stale.
23:55 ravana_2 Indeed, I'll to more testing to find out the reason. Thanks JoeJulian.

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