Camelia, the Perl 6 bug

IRC log for #gluster, 2013-09-21

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

All times shown according to UTC.

Time Nick Message
00:46 jag3773 joined #gluster
00:53 RileyBaker joined #gluster
00:58 RileyBaker joined #gluster
01:04 RileyBaker joined #gluster
01:13 bala joined #gluster
01:28 mooperd joined #gluster
01:40 _pol joined #gluster
02:42 zapotah joined #gluster
02:48 DV__ joined #gluster
03:12 sac joined #gluster
03:36 sac_ joined #gluster
03:36 sac`away` joined #gluster
03:37 sac`away joined #gluster
03:37 sac__ joined #gluster
03:45 neofob joined #gluster
03:53 kevein joined #gluster
03:54 sac`away joined #gluster
04:01 ericsean joined #gluster
04:04 ericsean Hi - sorry if this is common question. I setup gluster last night and it's all working great. however i noticed it has scrapped my previous nfs exports list. if i restart nfs-kernel-server i get my old exports back, but without the gluster volume i created. similarly if i restart glusterfs-server, i lose my old nfs exports but the gluster volume returns. Is there any simple way to configure them to co-exist? ubuntu on both hosts
04:10 JonathanD joined #gluster
04:15 Chocobo joined #gluster
04:17 mooperd joined #gluster
04:22 johnbot11 joined #gluster
04:28 Chocobo joined #gluster
04:39 Chocobo joined #gluster
04:45 Chocobo joined #gluster
04:55 Chocobo joined #gluster
04:55 Chocobo joined #gluster
05:11 Chocobo joined #gluster
05:14 _pol joined #gluster
05:27 johnbot11 joined #gluster
05:34 Guest89982 joined #gluster
05:51 davinder joined #gluster
05:56 Chocobo_ joined #gluster
06:15 davinder joined #gluster
06:19 satheesh joined #gluster
06:19 edong23_ joined #gluster
06:20 ThatGraemeGuy_ joined #gluster
06:20 d-fence joined #gluster
06:20 _Bryan_ joined #gluster
06:27 ProT-0-TypE joined #gluster
06:43 sgowda joined #gluster
06:55 bulde joined #gluster
07:13 bulde joined #gluster
07:15 ndevos ericsean: no, you can not. the nfs-server (and nlm and mountd) registers itself with rpcbind on localhost, it is not possible to register two nfs-servers there
07:29 bulde joined #gluster
07:50 bstr_ joined #gluster
07:50 ababu joined #gluster
07:54 dusmant joined #gluster
08:08 jporterfield joined #gluster
08:13 DataBeaver joined #gluster
10:07 mooperd joined #gluster
11:07 ericsean ndevos: so it is not possible to export nfs and run a glusterfs server on the same host?
11:08 ericsean ndevos: i think i meant to say export traditional nfs
11:19 Remco Correct
11:27 sticky_afk joined #gluster
11:27 stickyboy joined #gluster
11:54 davinder joined #gluster
12:40 ericsean ndevos,remco: thanks for the clarification
12:41 xavih joined #gluster
12:49 ndevos ericsean: well, you could, but when mounting you would need to pass the (on a Linux-NFS client) 'mountport' option to specify which nfs-server you want to talk to - thats not really user-friendly
12:50 ndevos and, the mountd port is dynamic, so you would need to configure that statically too
12:50 ericsean ndevos: that's a trade off i'm happy to make as i'm the sole nfs client on my network
12:51 ndevos ericsean: right, then you could go that route, I just have no idea how to configure that on ubuntu
12:51 ericsean ndevos: no worries, i'll check it out and report back if it's quick and easy. thanks for the tip
12:51 ndevos ericsean: see 'man 5 nfs' for the mountport option, it might explain the idea a little
12:52 ndevos ericsean: I do not think it is quick and easy, and if you dont know how mountd and nfs interact with each other, it probably is very confusing
12:53 ericsean ndevos: ok. the man page makes your rpcbind comment previously make more sense
12:54 ndevos ericsean: for testing, you should probably disable rpcbind or firewall port tcp+udp 111 from your nfs-client
12:55 ndevos ericsean: if you manage to mount the nfs-exports without rpcbind, you're half way already
12:55 ericsean ndevos: ok. does that mean the client won't need to run portmap locally either?
12:57 ndevos ericsean: I'm not sure, maybe you can specify a nlm-port (locking) as wel
12:58 ndevos ericsean: but, the local portmap==rpcbind does not really care whats on the nfs-server side
13:00 ndevos ericsean: well, actually, gluster comes with its own nlm, that might be more tricky (it would require a lnmport option for the nfs-mount too)
13:00 ndevos s/lnmport/nlmport/
13:00 uebera|| ndevos meant: “ericsean: well, actually, gluster comes with its own nlm, that might be more tricky (it would require a nlmport option for the nfs-mount too)”
13:00 glusterbot What ndevos meant to say was: ericsean: well, actually, gluster comes with its own nlm, that might be more tricky (it would require a nlmport option for the nfs-mount too)
13:03 bulde joined #gluster
13:07 zapotah joined #gluster
13:07 ericsean ndevos: The debian securing NFS page (https://wiki.debian.org/SecuringNFS) looks to be a good guide to do what you suggested
13:07 glusterbot Title: SecuringNFS - Debian Wiki (at wiki.debian.org)
13:08 ericsean ndevos: it seems a lot of customisation at the moment. might just use afp instead of nfs
13:09 ndevos ericsean: yeah, there are quite some options involved, and you can not pick a nlm-port per nfs-export/mount, it seems to be a (client) system-wide config: options lockd nlm_udpport=32768 nlm_tcpport=32768
13:10 ndevos (callback_tcpport is only used for NFSv4)
13:10 ndevos ericsean: oh, that would work! use the native-ubuntu NFSv4, and use NFSv3 for GlusterFS
13:10 ericsean ndevos: yeah. i'm just envisaging upgrading my server box at some point in the future and forgetting setting that all up
13:11 ndevos ericsean: NFSv4 does not use NLM/MOUNTD etc, so it is much easier to pass a single NFSv4 port with the mount options
13:12 ericsean ndevos: hmm, interesting . i don't think i've ever used nfsv4. will read up on that - cheers!
13:13 ndevos ericsean: difference in usage is minimal, just force the ubuntu-nfs-server to use v4 only, configure as usual (that would do it on RHEL/Fedora)
13:14 ndevos and, possibly disable registering the system-nfs-server to portmap/rpcbind so that GlusterFS-nfs can register without issues
13:24 xavih joined #gluster
13:33 ericsean ndevos: success!  now going to try mounting glusterfs and nfs side by side on client
13:36 ericsean ndevos: working perfectly. thanks so much for your help - would have been on google all night!!
13:36 ndevos ericsean: cool, nice to hear it works :)
13:40 samppah any thoughts about gluster monitoring? i'm currectly using munin to monitor disk and network performance
13:46 jag3773 joined #gluster
13:46 bstr_ joined #gluster
13:59 ndevos samppah: I hope to find some time to try Glubix (https://forge.gluster.org/glubix) in the next few weeks, it looks promising to me
13:59 glusterbot Title: Glubix - Gluster Community Forge (at forge.gluster.org)
13:59 * ndevos thinks that name sounds like a character from Asterix & Obelix
14:29 samppah ndevos: heh, you are correct.. it sounds like friend of Asterix :)
14:29 samppah i have been testing zabbix too so i think that might be worth a try
14:30 shylesh joined #gluster
14:43 johnbot11 joined #gluster
14:45 johnbot11 joined #gluster
14:59 vimal joined #gluster
15:01 ProT-0-TypE joined #gluster
15:13 ProT-0-TypE joined #gluster
15:16 ThatGraemeGuy joined #gluster
16:08 dusmant joined #gluster
17:28 zapotah joined #gluster
17:54 _pol joined #gluster
18:08 johnbot11 joined #gluster
18:08 johnbot11 joined #gluster
18:53 fyxim joined #gluster
19:33 johnbot11 joined #gluster
19:56 ProT-0-TypE joined #gluster
20:02 lanning joined #gluster
20:06 johnmwilliams joined #gluster
21:11 mooperd joined #gluster
21:37 johnbot11 joined #gluster
22:19 johnbot11 joined #gluster
23:46 TDJACR joined #gluster
23:50 johnbot11 joined #gluster

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