Perl 6 - the future is here, just unevenly distributed

IRC log for #gluster, 2017-10-30

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

All times shown according to UTC.

Time Nick Message
00:09 bEsTiAn joined #gluster
00:33 daMaestro joined #gluster
00:51 gbox protoporpoise: failover happens automatically using the fuse client.  You can mount from any host
00:52 protoporpoise Yeah but if you're mount point on all your clients is host01, host01 dies - how does the client that then starts and tries to mount host01 mount host02 instead?
00:56 gbox protoporpoise: the logic is in the fuse client.  Much of the "guts" of gluster exists in the client
00:57 protoporpoise But the fuse client can only be passed one server hostname to attempt the mount, so if that server isn't available and the client isn't connected - I can't see any way for it to try another host in that same pool
00:57 gbox protoporpoise: it's the same way that you get info from "gluster volume info <volume name>".  If you want to know the technical details, it's all in /var/run/gluster/* files
00:58 protoporpoise but if it doesn't have that information because it hasn't connected to the pool...
00:58 protoporpoise e.g. we spin up clients many times a day, if one is spun up and host01 isn't available it fails to connect
00:58 gbox protoporpoise: OK I see what you mean. Like if you have host1 in /etc/fstab, and then need to mount when host1 is down.  Good question!
00:58 protoporpoise yep you got it
00:59 protoporpoise The result is that the client connection system doesn't scale well when used with modern hosting environments
01:03 gbox backupvolfile
01:04 gbox There's a backupvolfile option for the mount command that allow for failover in case the primary node is down
01:05 protoporpoise ok interesting, I'll RTFM on that
01:05 protoporpoise thanks so much for pointing me to it
01:05 gbox protoporpoise:  No problem, gave me that chance to learn it too!
01:06 protoporpoise hahaha, often the way hey, it looks like you can only specify one host, but it's better than no alternative
01:07 gbox protoporpoise:  It says "Provide list of backup volfile servers" but the format is not clear.  Comma or space delimited?
01:08 protoporpoise oh in the latest version yes I see that, I was on the older versions doco page
01:08 protoporpoise Link for IRC logs / anyone else interested: https://access.redhat.com/documentation/en-us/red_hat_gluster_storage/3.3/html-single/administration_guide/#sect-Mounting_Red_Hat_Storage_Volumes
01:08 glusterbot Title: Administration Guide - Red Hat Customer Portal (at access.redhat.com)
01:09 bEsTiAn joined #gluster
01:11 gbox protoporpoise:Yeah looks like they worked on it recently, and those docs are better than before too
01:11 protoporpoise Yeah - I probably asked too many 'webscale' questions lol
01:12 gbox backup-volfile-servers colon-delimited just the hostname
01:12 protoporpoise We're using GlusterFS for our new storage for our Kubernetes Docker clusters
01:12 gbox yeah you need more keywords in there
01:13 gbox buzzwords.  So what config management do you use for all that?
01:13 protoporpoise For kubes?
01:13 protoporpoise We automate everything with Puppet
01:14 gbox yeah webscale for sure
01:14 protoporpoise ;)
01:14 protoporpoise diverse modern web app hosting calls for a lot of automation
01:14 protoporpoise helps use reduce DR times greatly and make sure things are repeatable
01:14 protoporpoise especially when you have a _very_ small budget
01:15 protoporpoise (non-profit, charitable organisation)
02:10 bEsTiAn joined #gluster
02:16 plarsen joined #gluster
02:26 daMaestro joined #gluster
02:36 ahino joined #gluster
02:55 ilbot3 joined #gluster
02:55 Topic for #gluster is now Gluster Community - https://www.gluster.org | Documentation - https://gluster.readthedocs.io/en/latest/ | Patches - https://review.gluster.org/ | Developers go to #gluster-dev | Channel Logs - https://botbot.me/freenode/gluster/ & http://irclog.perlgeek.de/gluster/
03:02 prasanth joined #gluster
03:15 gyadav_ joined #gluster
03:20 protoporpoise another question: I'm struggling to find documentation in regards as to what should be backed up (other than data) on gluster hosts, for example - is the information from `gluster get-state` enough to somehow be restored in a disaster event?
03:26 protoporpoise (without using thin LVM snapshots)
03:39 psony joined #gluster
04:06 kramdoss__ joined #gluster
04:11 bEsTiAn joined #gluster
04:13 msvbhat joined #gluster
04:30 ronrib joined #gluster
04:39 sunny joined #gluster
05:04 protoporpoise head off for the day but will leave IRC open - interested to know who's using various TCP congestion control algorithms with Glusters, Cubic (default in Linux), HTCP (I've seen recommended for use with Gluster), something else (reno perhaps?)
05:11 bEsTiAn joined #gluster
05:15 hgowtham joined #gluster
05:20 karthik_us joined #gluster
05:29 ndarshan joined #gluster
05:33 Saravanakmr joined #gluster
05:37 prasanth joined #gluster
05:44 skumar joined #gluster
05:59 skumar joined #gluster
06:08 buvanesh_kumar joined #gluster
06:11 bEsTiAn joined #gluster
06:19 ronrib_ joined #gluster
06:29 xavih joined #gluster
06:35 gyadav_ joined #gluster
06:50 kramdoss__ joined #gluster
06:50 skoduri joined #gluster
07:12 bEsTiAn joined #gluster
07:14 christianm joined #gluster
07:15 skumar joined #gluster
07:19 mbukatov joined #gluster
07:22 kdhananjay joined #gluster
07:25 skumar_ joined #gluster
07:26 gyadav__ joined #gluster
07:33 gyadav_ joined #gluster
07:35 jtux joined #gluster
07:37 Teraii joined #gluster
07:44 buvanesh_kumar joined #gluster
07:47 [diablo] joined #gluster
08:02 omie888777 joined #gluster
08:14 fsimonce joined #gluster
08:21 ivan_rossi joined #gluster
08:29 rouven joined #gluster
08:42 sanoj joined #gluster
08:57 _KaszpiR_ joined #gluster
08:57 kramdoss__ joined #gluster
09:07 prasanth joined #gluster
09:13 christianm joined #gluster
09:14 Saravanakmr joined #gluster
09:22 DavidRama joined #gluster
09:24 DavidRama hi all, I'm trying to use Glusterfs from my kubernetes. I got issues because the gluster nodes are behind a NAT and thus the Kubernetes tries to talk to the non Nated IP and fails as it has no route to reach the cluster. Are there any docs/links about this kind of setup ? Thanks a lot
09:28 misc DavidRama: I think you need to do some kind of port forwarding :/
09:29 DavidRama misc: the port forwarding can't help with the routing I guess. Glusterfs can't work behind a NAT ?
09:30 sanoj joined #gluster
09:35 misc DavidRama: so the nat is between what and what ?
09:35 ahino joined #gluster
09:35 DavidRama the k8s cluster and the gluster cluster
09:36 DavidRama k8s is in my public space
09:37 DavidRama yhen putting the NAT'ed IP I get this messages: [MSGID: 106301] [glusterd-syncop.c:1297:gd_stage_op_phase] 0-management: Staging of operation 'Volume Create' failed on localhost : Host 10.0.45.146 is not in ' Peer in Cluster' state
09:37 DavidRama from the glusterfs
09:40 foster joined #gluster
09:41 misc so, you have a gluster cluster, whose members are on a network, and behind a nat, and you want kubernete to speak to it
09:41 misc mhh, yeah, I think you need them in the same network
09:50 bfoster joined #gluster
10:07 christianm joined #gluster
10:11 ThHirsch joined #gluster
10:51 jkroon joined #gluster
11:00 skumar joined #gluster
11:01 [fre] joined #gluster
11:01 [fre] hi Guys.
11:02 [fre] Just a hopefully quick little question.
11:04 [fre] Erroneously one of our volumes got created with only the hostname, no domain specified. While it works for inter-node-communication, it does not for client-connections. Is there a way to change the config without rebuilding the volume?
11:22 MrAbaddon joined #gluster
11:42 DV joined #gluster
12:05 marlinc joined #gluster
12:06 plarsen joined #gluster
12:09 phlogistonjohn joined #gluster
12:21 skumar_ joined #gluster
12:22 arpu joined #gluster
12:25 christianm joined #gluster
12:31 skumar_ joined #gluster
12:35 msvbhat joined #gluster
12:36 buvanesh_kumar joined #gluster
12:40 rwheeler joined #gluster
12:43 kramdoss__ joined #gluster
12:52 skumar__ joined #gluster
13:03 guhcampos joined #gluster
13:06 guhcampos_ joined #gluster
13:33 dominicpg joined #gluster
13:34 jkroon joined #gluster
13:40 hmamtora joined #gluster
13:46 brayo joined #gluster
13:57 phlogistonjohn joined #gluster
14:07 gyadav_ joined #gluster
14:12 farhorizon joined #gluster
14:22 matt__ joined #gluster
14:23 matt__ I'm trying to get gluster setup with ganesha. I have gluster setup right now with a volume. The volume is started. Trying to export it now using nfs-ganesha.
14:23 matt__ Not sure what I am doing wrong. This is on CentOS 7, gluster 3.12
14:24 matt__ I've tried to do ganesha.enable, but it says that the option doesnt exist
14:24 matt__ trying to follow this: http://docs.gluster.org/en/latest/Administrator%20Guide/NFS-Ganesha%20GlusterFS%20Integration/
14:24 glusterbot Title: Configuring NFS-Ganesha server - Gluster Docs (at docs.gluster.org)
14:34 d4n13L joined #gluster
14:35 DJClean joined #gluster
14:42 Klas matt__: ganesha is deprecated or maybe even removed, unless I'm mistaken
14:42 Klas there's a new shiny way to do it
14:43 Klas (haven't tried either of them and I can't remember the name of the new solution)
14:45 ThHirsch joined #gluster
14:48 matt__ Klas: Would you happen to know where to find that. I think I found the issue. The package name was different and I didn't realize that it didnt install
14:48 Klas no idea, sorry
14:48 Klas never played with nfs HA in gluster
14:49 matt__ Klas: but if theres a newer better way, I'd like to know. I'm trying to see if NFS would be faster since it has caching
14:49 Klas nfs non-ha is definitely faster
14:49 Klas about twice as fast in my experience
14:50 matt__ Klas: yeah, im trying to get some redundancy, I don't really mind not having HA, but using fuse, it's incredibly slow for what I'm trying to do
14:51 MrAbaddon joined #gluster
14:51 farhorizon joined #gluster
14:53 jirib joined #gluster
14:53 jirib hi all
14:53 jirib does mounting glusters volume work over ipv6?
14:55 jirib https://pastebin.com/chY31LgP
14:55 glusterbot Please use http://fpaste.org or http://paste.ubuntu.com/ . pb has too many ads. Say @paste in channel for info about paste utils.
14:58 kpease joined #gluster
14:59 kpease_ joined #gluster
15:04 wushudoin joined #gluster
15:08 DavidRama left #gluster
15:17 matt__ Klas: just fyi, i think ganesha is built-in now. No need to enable it anymore from what I can tell. Yes, it's much much faster. Especially for smaller files
15:21 Klas ah
15:21 Klas nice to know =)
15:22 sunny joined #gluster
15:38 vbellur joined #gluster
15:44 snave joined #gluster
16:01 MrAbaddon joined #gluster
16:06 guhcampos_ joined #gluster
16:35 ThHirsch joined #gluster
16:37 farhorizon joined #gluster
16:40 skoduri joined #gluster
16:43 shyam joined #gluster
16:48 buvanesh_kumar joined #gluster
16:48 guhcampos joined #gluster
16:50 vbellur joined #gluster
16:52 ^andrea^ joined #gluster
16:56 shruti joined #gluster
16:59 BitByteNybble110 joined #gluster
17:13 ThHirsch joined #gluster
17:23 prasanth joined #gluster
17:27 farhorizon joined #gluster
17:32 ThHirsch joined #gluster
17:33 vbellur joined #gluster
17:48 farhorizon joined #gluster
17:55 ThHirsch joined #gluster
17:56 ivan_rossi left #gluster
18:11 ThHirsch1 joined #gluster
18:12 MrAbaddon joined #gluster
18:31 farhorizon joined #gluster
18:46 BitByteNybble110 joined #gluster
18:55 ThHirsch joined #gluster
19:14 _KaszpiR_ joined #gluster
19:44 Jacob843 joined #gluster
19:54 vbellur joined #gluster
20:19 vbellur joined #gluster
20:24 social joined #gluster
20:29 farhorizon joined #gluster
20:37 vbellur joined #gluster
20:39 vbellur joined #gluster
20:39 vbellur1 joined #gluster
20:41 vbellur joined #gluster
20:42 vbellur1 joined #gluster
20:56 farhorizon joined #gluster
21:07 DV joined #gluster
21:21 pioto joined #gluster
21:51 xniega joined #gluster
22:38 hmamtora joined #gluster

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