Camelia, the Perl 6 bug

IRC log for #gluster, 2013-10-04

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

All times shown according to UTC.

Time Nick Message
00:06 nasso joined #gluster
00:12 sprachgenerator joined #gluster
00:24 bdperkin_gone joined #gluster
00:38 vpshastry joined #gluster
00:44 StarBeast joined #gluster
00:52 dtyarnell joined #gluster
00:55 bala joined #gluster
01:16 sprachgenerator joined #gluster
01:19 bala joined #gluster
01:22 harish joined #gluster
01:28 DV joined #gluster
01:53 eryc joined #gluster
01:53 eryc joined #gluster
02:24 bdperkin joined #gluster
02:26 satheesh1 joined #gluster
02:26 kanagaraj joined #gluster
02:51 kshlm joined #gluster
02:58 theogonica joined #gluster
03:03 theogonica joined #gluster
03:05 shubhendu joined #gluster
03:22 dusmant joined #gluster
03:30 lpabon joined #gluster
03:30 sgowda joined #gluster
03:31 davinder joined #gluster
03:36 shyam joined #gluster
03:45 itisravi joined #gluster
03:50 shubhendu joined #gluster
04:05 anands joined #gluster
04:09 shylesh joined #gluster
04:14 shruti joined #gluster
04:37 rjoseph joined #gluster
04:38 bala joined #gluster
04:38 shubhendu joined #gluster
04:47 glusterbot New news from newglusterbugs: [Bug 1010874] Dist-geo-rep : geo-rep config log-level option takes invalid values and makes geo-rep status defunct <http://goo.gl/om4qdi>
04:56 bala joined #gluster
04:58 psharma joined #gluster
05:02 lalatenduM joined #gluster
05:08 shyam joined #gluster
05:09 CheRi joined #gluster
05:13 raghu joined #gluster
05:19 aravindavk joined #gluster
05:22 Guest34797 joined #gluster
05:34 anands joined #gluster
05:37 rastar joined #gluster
05:43 ppai joined #gluster
05:46 nshaikh joined #gluster
05:52 an joined #gluster
05:53 RedShift joined #gluster
05:54 RedShift hi all
05:54 RedShift imagine the following scenario, a gluster setup hosting virtual machines (continous I/O). I have node1 and node2, node1 is active and node2 is backup. node1 & node2 need servicing which require them to be taken offline.
05:54 RedShift I take node1 offline, the failover mechanism puts node2 active, thus all writes proceed on node2.
05:54 RedShift after node1 has been brought back online, all changes replicate to node1 from node2
05:54 RedShift but how can I tell if the replication has been completed and I can take node2 offline? Since a lot of virtual machines can cause a lot of IO, there might not be a point where the number of files to be healed is 0
05:54 RedShift if I take node2 offline too soon, it will have pending healing data to be transferred to node1 and node1 will already have new data because failover will have put node1 in charge because node2 was offline
05:54 RedShift thus - a split brain?
06:00 DV joined #gluster
06:01 samppah RedShift: you can use command "gluster volume heal volName info" to follow that there isn't replication pending
06:04 RedShift samppah yes but what if my virtual machines are so busy all the time, there's always a file in replication state?
06:10 samppah RedShift: if everything is fine files will usually just pop in that list and then vanish
06:10 RedShift even when high IO is occuring?
06:10 samppah if there is something wrong they will show up for longer time
06:10 samppah yes.. at least that's what i have noticed
06:11 bulde joined #gluster
06:13 rgustafs joined #gluster
06:18 satheesh1 joined #gluster
06:18 vimal joined #gluster
06:29 bulde joined #gluster
06:32 bulde1 joined #gluster
06:32 jtux joined #gluster
06:33 anands joined #gluster
06:37 davinder joined #gluster
06:40 shyam joined #gluster
06:40 samppah RedShift: i think you are correct.. files seems to stay on that list as long as there is IO going on so i think it may cause problems if they are IO heavy
06:42 samppah i just did quick tests with dd and fio
06:48 sgowda joined #gluster
06:49 DV joined #gluster
06:50 purpleidea PatNarciso: hey, i read your google docs setup. Despite your comments I think that using gluster in a geo-replicated setup is probably not recommended. You should at least measure the latency, and run a test cluster (eg: vm's) first before doing this. Keep in mind that split brain will be a significant if not unsolveable challenge. The only out I can think of is if you're doing one way replication. In that case, I still am not sure gluster is t
06:51 purpleidea (and to clarify, the way i understand this is you're trying to do regular gluster replication, just across two different states on a 10mbit/sec symmetrical connection.
06:51 purpleidea )
06:52 rjoseph joined #gluster
06:53 ekuric joined #gluster
06:53 ctria joined #gluster
06:54 Maxence_ joined #gluster
06:54 ricky-ticky joined #gluster
06:55 rastar joined #gluster
06:57 satheesh2 joined #gluster
07:00 lalatenduM_ joined #gluster
07:03 ekuric left #gluster
07:04 shyam joined #gluster
07:05 ngoswami joined #gluster
07:12 ekuric joined #gluster
07:38 andreask joined #gluster
07:46 lalatenduM_ joined #gluster
07:47 rjoseph joined #gluster
07:56 mgebbe_ joined #gluster
08:06 rastar joined #gluster
08:10 shubhendu joined #gluster
08:11 RameshN joined #gluster
08:12 pkoro joined #gluster
08:17 shyam1 joined #gluster
08:18 glusterbot New news from newglusterbugs: [Bug 990028] enable gfid to path conversion <http://goo.gl/1HwiQc> || [Bug 969461] RFE: Quota fixes <http://goo.gl/XFSM4>
08:20 sgowda joined #gluster
08:30 aravindavk joined #gluster
08:33 ppai joined #gluster
08:33 pkoro joined #gluster
08:38 StarBeast joined #gluster
08:54 Maxence joined #gluster
08:59 rastar joined #gluster
09:09 ngoswami joined #gluster
09:11 satheesh joined #gluster
09:14 harish_ joined #gluster
09:14 ekuric joined #gluster
09:24 jcsp joined #gluster
09:25 psharma joined #gluster
09:37 samsamm joined #gluster
09:47 keytab Hello, Currently I have a glusterfs server with 2 briks and also I have a client that mount gluster partition. If i will down server number one mi client can't acces to mount partition about 1 minut, late respond corretly, it is normal?
09:48 samppah keytab: yes, there is option network.ping-timeout which is 42 sec by default.. it waits for that period of time before it declares server dead
09:49 samppah IO is halted for that time period
09:54 khushildep_ joined #gluster
09:54 RameshN joined #gluster
09:56 badone joined #gluster
09:57 keytab thanks samppah
09:59 keytab and the last question, glusterfs write in logs with 2 hours less but my other logs write well
10:03 vshankar joined #gluster
10:04 RameshN joined #gluster
10:08 lkthomas joined #gluster
10:08 samppah keytab: yeah, logs use  UTC for some reason
10:08 samppah if i understood correctly what you mean :)
10:09 keytab You understand me corrently
10:25 davinder joined #gluster
10:27 dusmant joined #gluster
10:59 rjoseph joined #gluster
10:59 sgowda joined #gluster
11:04 ngoswami joined #gluster
11:20 ppai joined #gluster
11:33 rastar joined #gluster
11:38 harish_ joined #gluster
11:44 khushildep joined #gluster
11:48 social_ how thrustworthy is 3.4.1 and gluster volume replace-brick?
11:57 psharma joined #gluster
11:58 andreask joined #gluster
12:05 ninkotech joined #gluster
12:05 ninkotech_ joined #gluster
12:09 itisravi_ joined #gluster
12:21 thanasisk joined #gluster
12:21 thanasisk split brain found, aborting selfheal of / - how can i fix this?
12:22 thanasisk do I need to rebalance?
12:24 Remco @splitbrain
12:24 glusterbot Remco: To heal split-brain in 3.3+, see http://goo.gl/FPFUX .
12:27 thanasisk Remco, im on 3.2.7-3 on Debian
12:27 thanasisk heal does not get recognized as a keyword
12:28 Remco That's more complex then I think
12:28 Remco Same version for me, but I'm thinking of upgrading it
12:29 Remco And never seen it happen yet, but then I only have a test setup at the moment
12:29 thanasisk any tips on how to identify split-brain files on 3.2.X?
12:30 Remco I have none :(
12:30 social_ anyone with direct link to gluster bugzilla?
12:30 ndevos social_: see file a bug
12:30 glusterbot http://goo.gl/UUuCq
12:31 ninkotech joined #gluster
12:33 B21956 joined #gluster
12:35 ingard__ !oldbug 1188
12:35 glusterbot Bug http://goo.gl/kFkthH high, low, 3.1.1, pkarampu, CLOSED CURRENTRELEASE, 3.0.5 client crash - afr_set_split_brain
12:40 shyam1 left #gluster
12:43 abradley joined #gluster
12:49 glusterbot New news from newglusterbugs: [Bug 1015517] gluster crash during replace brick <http://goo.gl/J16RUq> || [Bug 1011662] threads created by gluster should block signals which are not used by gluster itself <http://goo.gl/q9M1z5>
13:01 bennyturns joined #gluster
13:01 thanasisk joined #gluster
13:01 thanasisk how can I rotate the logs in the client?
13:17 nasso joined #gluster
13:22 dtyarnell joined #gluster
13:29 bala joined #gluster
13:36 bulde joined #gluster
13:41 abradley I'm a gluster newb looking to setup on ubu12server
13:42 abradley Is the 3.4 install any different than the 3.2 install detailed here? http://www.howtoforge.com/high-availability-stor​age-with-glusterfs-3.2.x-on-ubuntu-12.04-automat​ic-file-replication-across-two-storage-servers
13:42 glusterbot <http://goo.gl/p462Z> (at www.howtoforge.com)
13:44 l0uis abradley: after a cursory look, appears to be the same
13:44 abradley cool, thanks for your input
13:44 l0uis abradley: add the ppa, install the package, peer probe, create volume
13:44 l0uis abradley: its sort of embarrasingly simple, to be honest, given how powerful gluster is :)
13:44 abradley thanks for the help
13:45 l0uis abradley: if you do this for a living, use the George Costanza approach to looking busy since gluster is so easy, so your boss doesn't get suspicious: http://youtu.be/yd9ma2UVLHM
13:45 glusterbot Title: How To: Look like youre busy. The George Costanza way - YouTube (at youtu.be)
13:46 abradley haha
13:49 davinder joined #gluster
13:54 bugs_ joined #gluster
13:55 abradley l0uis, you mentioned adding the ppa. I dont's ee where this guide mentions that: http://www.howtoforge.com/high-availability-stor​age-with-glusterfs-3.2.x-on-ubuntu-12.04-automat​ic-file-replication-across-two-storage-servers
13:55 glusterbot <http://goo.gl/p462Z> (at www.howtoforge.com)
13:59 kaptk2 joined #gluster
14:00 abradley a second question, where can one find the GUI shown here? http://www.youtube.com/watch?v=pJEsB-gUdKk
14:00 glusterbot Title: How to Install Gluster Storage Platform - YouTube (at www.youtube.com)
14:05 khushildep joined #gluster
14:11 l0uis abradley: ubuntu ships w/ an old gluster. you want to run 3.4.1 right?
14:12 l0uis abradley: if so, add the ppa from semiosis
14:12 abradley indeed
14:12 l0uis let's see if this works
14:12 l0uis ,,ppa
14:13 abradley I'll add that step as "step 1" yes?
14:13 l0uis @ppa
14:13 glusterbot l0uis: The official glusterfs packages for Ubuntu are available here: 3.3 stable: http://goo.gl/7ZTNY -- 3.4 stable: http://goo.gl/u33hy
14:14 l0uis ah there we go :)
14:14 l0uis sorry, my glusterbot foo is lacking
14:14 l0uis yeah, first step is to add the ppa. then update. the install 3.4 (or upgrade what you have)
14:14 abradley thanks a lot
14:16 anands joined #gluster
14:20 samsamm joined #gluster
14:27 sprachgenerator joined #gluster
14:29 lpabon joined #gluster
14:34 failshell joined #gluster
14:34 failshell hello. im trying to write to a samba exported volume on 3.4.0. but im gettting this error: 0-test-client-0: remote operation failed: Permission denied. Path:
14:35 failshell i followed the instructions here: https://access.redhat.com/site/documentat​ion//en-US/Red_Hat_Storage/2.1/html/Admin​istration_Guide/sect-Administration_Guide​-GlusterFS_Client-CIFS.html#idp3948496
14:35 glusterbot <http://goo.gl/C0ONZI> (at access.redhat.com)
14:59 failshell E [afr-common.c:2022:afr_lookup_done] 0-test-replicate-0: /..: No gfid present
14:59 failshell what does that mean?
15:03 ndk joined #gluster
15:07 phox failshell: is that as in like, csh/tcsh?
15:07 phox I would assume that means the brick does not have a gfid xattr set for its root
15:08 dtyarnell joined #gluster
15:09 phox although oddly one of mine has neither... I think
15:10 * phox shrugs
15:10 bala joined #gluster
15:11 abradley I added ppa:semiosis/ubuntu-glusterfs-3.4 and then apt-get install glusterfs-server
15:11 abradley but it installed v3.2.5 and not 3.4
15:12 failshell phox: doesnt seem to cause any issues. i guess  ill just ignore it
15:13 phox huh, weird
15:14 phox hey people.. is there any actual documentation that specifies what exactly nobootwait does?
15:14 phox might be a feature I was looking for, but might also not quite be
15:15 abradley "GlusterFS comes with ABSOLUTELY NO WARRANTY.
15:15 abradley More info: https://launchpad.net/~semiosis​/+archive/ubuntu-glusterfs-3.4
15:15 abradley Press [ENTER] to continue or ctrl-c to cancel adding it
15:15 abradley " I hit enter and the ppa was added successfully but apt-get install glusterfs-server installed 3.2.5
15:15 glusterbot <http://goo.gl/u33hy> (at launchpad.net)
15:16 Guest34797 joined #gluster
15:16 phox probably because of what dist it's set to within the package
15:16 phox i.e. it might be unstable or experimental
15:17 phox which is kinda dumb, bceause TBH 3.2 is probably less "stable" than newer, but... welcome to distros
15:18 abradley how would one go about installing 3.4 then?
15:18 phox probably apt-get -t <whatever> install ....
15:18 phox BUT
15:18 phox as in our case, you may want to consider installing ONLY the deps you NEED to be from experimental or whatever
15:18 phox or unstable or...
15:19 phox by using e.g. /etc/apt/apt-preferences.d/.... (I think that's the right path)
15:19 abradley Oh is 3.4.1 experimental/unstable?
15:19 phox most people just use apt-get -t whatever
15:19 phox it might be
15:19 phox apt-cache showpkg whatever
15:22 abradley is it possible that I must specify this custom repository when apt-get?
15:23 abradley apt-get install glusterfs-server semiosis.... ?
15:29 abradley forgive me, but what does -t do after apt-get ?
15:31 abradley apt-get help doesn't show -t
15:31 LoudNoises joined #gluster
15:34 fuzzy_id joined #gluster
15:34 fuzzy_id hi
15:34 glusterbot fuzzy_id: Despite the fact that friendly greetings are nice, please ask your question. Carefully identify your problem in such a way that when a volunteer has a few minutes, they can offer you a potential solution. These are volunteers, so be patient. Answers may come in a few minutes, or may take hours. If you're still in the channel, someone will eventually offer an answer.
15:36 semiosis abradley: ???
15:37 abradley semiosis, I didn't realize you were a user in here. That was not directed at you.
15:37 abradley are you associated with the ppa?
15:38 abradley if so, how would I go about installing glusterfs 3.4? I added your ppa but apt-get install glusterfs-server installed 3.2.5
15:38 l0uis abradley: did you update after you added the repository?
15:38 abradley aHA! I did NOTG
15:38 l0uis abradley: that there is your problem then :)
15:38 abradley Thank you so much
15:38 semiosis thx l0uis
15:39 semiosis add-apt-repository; apt-get update; apt-get install
15:39 fuzzy_id well, as we are at it: is ppa:semiosis/ubuntu-glusterfs-3.4 ready for production? I'm on ubuntu 12.04 and wondering if it is preferrable to install GlusterFS 3.4 over 3.2 from the repos…
15:39 semiosis abradley: you should probably remove the old packages (and kill any running processes) before you start with the ppa, just to keep things clean
15:40 l0uis fuzzy_id: i think 3.4 is very much preferred
15:40 l0uis fuzzy_id: i'm using it in production.
15:40 semiosis fuzzy_id: what do you mean?
15:42 fuzzy_id semiosis: nothing specific, just wondering how stable that thing is, if there are much bugs to fix, if packaging interacts nicely with the rest of ubuntus repo, blahblah…
15:43 semiosis how stable is anything? :)
15:43 semiosis use the latest GA release of glusterfs (currently 3.4.1) and report bugs if you find any
15:43 fuzzy_id ok
15:45 failshell im using RHS 2.1 (3.4.0) and its much stabler than the previous 3.3.0 build
15:46 fuzzy_id is it possible to activate master to master encryption in 3.4?
15:47 fuzzy_id i've seen `server.ssl' and `client.ssl', i guess these two are for client-server interaction, right?
15:48 fuzzy_id and then there seems to be `transport.socket.ssl-enabled'…
15:48 fuzzy_id https://lists.gnu.org/archive/html/​gluster-devel/2012-10/msg00103.html
15:48 glusterbot <http://goo.gl/vEZDWh> (at lists.gnu.org)
15:55 semiosis fuzzy_id: last i heard multi master geo-replication will be in 3.7
15:55 semiosis no idea how accurate that is, but i heard it
16:02 johnmark fuzzy_id: encryption between servers?
16:02 fuzzy_id i don't get this
16:02 fuzzy_id johnmark: yep
16:03 johnmark fuzzy_id: not sure. Our in-flight encryption is between GlusterFS client and data pool
16:03 johnmark we're workign on at-rest encryption for the data pool itself, but I don't know when it will be ready
16:05 fuzzy_id semiosis: does that mean that glusterfs with master to master replication should actually only be used in LANs?
16:05 semiosis two kinds of replication... the regular, unqualified "replication" (called AFR) and then geo-replication
16:06 semiosis yes AFR works best on a LAN
16:06 semiosis it *can* work over high latency links, but you probably won't be happy with the performance
16:06 semiosis that's what geo-rep is for, but currently that is one-way master to slave
16:07 fuzzy_id hmm
16:07 fuzzy_id what do you mean with high latency?
16:07 fuzzy_id latency > 10ms?
16:08 fuzzy_id i just benchmarked glusterfs with a latency of about 8.6ms:
16:08 fuzzy_id http://blog.ilexius.de/blog-post​/2013/09/27/glusterfs-benchmark/
16:08 glusterbot <http://goo.gl/4I6Aq9> (at blog.ilexius.de)
16:08 fuzzy_id and performance seemed to be allright
16:09 semiosis high is subjective
16:09 fuzzy_id ok, granted
16:09 semiosis "too high for your purposes"
16:09 hagarth joined #gluster
16:09 semiosis we like ,,(joe's performance metric) around here
16:09 glusterbot nobody complains.
16:10 fuzzy_id but if throughput is the only issue i'd give it a shot…
16:10 semiosis but quite often people complain about running AFR over WAN links
16:11 DV joined #gluster
16:30 Mo__ joined #gluster
16:38 l0uis semiosis: is that because of latency, or because of the bandwidth hit in general, since its probably a lot more noticable that you're running at only half speed (for replicas=2)
16:39 wushudoin joined #gluster
16:42 jporterfield joined #gluster
16:43 abradley I have glusterfs-server setup on two machines and they are connected to one another. Each machine has one hard drive with two partitions: one (~4gb) running ubu12server and glusterfs-server and the other (~140gb) that is unpartitioned. I have created a replicated volume "proxmox". How do I add this unpartitioned space on each server to the proxmox volume?
16:43 semiosis l0uis: some of one, some of the other
16:43 phox semiosis: maybe you know this.  looking for documentation on exactly what nobootwait does.
16:44 phox seems most of the switches and flags and stuff are all OMG SECRET which gets kind old =/
16:44 abradley clarify: each server has one ~160gb hard drive with 4 gb on the OS partition and the rest unpartitioned
16:44 semiosis phox: see man fstab
16:44 phox semiosis: oh, not specific to gluster?
16:44 * phox thought it was
16:44 semiosis phox: it's ubuntu
16:44 phox -oh-
16:44 semiosis http://manpages.ubuntu.com/manpa​ges/precise/en/man5/fstab.5.html
16:44 glusterbot <http://goo.gl/N1eLJy> (at manpages.ubuntu.com)
16:44 phox so kinda like _netdev
16:44 phox or wahtever.
16:45 phox ok.  thought it was some sneaky gluster thing.  thanks.
16:45 l0uis abradley: you have to partition the gluster brick
16:45 semiosis well they're both used for remote mounts, but they mean different things because the boot processes are different
16:45 l0uis abradley: and put a file system on it
16:45 l0uis abradley: then add the brick to gluster
16:45 abradley do you recommend fdisk or some other app?
16:45 abradley -over ssh
16:45 l0uis abradley: xfs is recommended for the fs.
16:45 abradley ok, good info, I was thinking ext4
16:45 l0uis abradley: your tool of choice doesnt matter as far as i know
16:45 l0uis abradley: you should read the docs :)
16:45 semiosis ,,(rtfm)
16:45 glusterbot Read the fairly-adequate manual at http://goo.gl/E3Jis
16:46 l0uis they certainly aren't perfect but cover the basics
16:46 l0uis abradley: notable the quickstart: http://www.gluster.org/community/d​ocumentation/index.php/QuickStart
16:46 glusterbot <http://goo.gl/OEzZn> (at www.gluster.org)
16:47 l0uis @rtfm
16:47 glusterbot l0uis: Read the fairly-adequate manual at http://goo.gl/E3Jis
16:47 l0uis oh i see.. @.. informs glusterbot to tell me, whereas ,,(..) tells the channel
16:47 glusterbot l0uis: Error: No factoid matches that key.
16:48 semiosis l0uis: ,,(rtfm) can also be sent to a person
16:48 glusterbot l0uis: Read the fairly-adequate manual at http://goo.gl/E3Jis
16:48 semiosis glusterbot: ,,(rtfm) or a bot
16:48 semiosis or not
16:49 l0uis semiosis: @rtfm this too?
16:49 l0uis nope
16:50 abradley primary partition vs logical: is one better than the other for gluster?
16:51 l0uis abradley: i dont think it matters. when in doubt, i alway spick primary if i can
16:52 bulde joined #gluster
16:58 abradley I've setup a new partition in ubuserver and now I wish to format
16:58 abradley http://pastebin.com/Yg0CqSiL
16:58 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.
16:59 abradley http://paste.ubuntu.com/6193017/
16:59 glusterbot Title: Ubuntu Pastebin (at paste.ubuntu.com)
16:59 abradley I created a new partition with fdisk
16:59 abradley now I have 4 different options so I'm not sure which one is the one I just created. It hsould be the largest
17:00 abradley sda2 and sda5 look extremely similar
17:03 abradley I chose sda2 and I get this error: http://paste.ubuntu.com/6193032/
17:03 glusterbot Title: Ubuntu Pastebin (at paste.ubuntu.com)
17:04 l0uis abradley: did you read the quickstart?
17:05 abradley xfsprogs, got it
17:05 abradley thanks
17:05 abradley would sda2 be the correct choice?
17:05 abradley instead of 5?
17:06 abradley sda2 is "extended" and sda5 is linux lvm
17:06 l0uis abradley: i have no idea which is correct for your server
17:06 l0uis but either way, you really should read the quickstart becuase you don't point gluster at a device
17:06 l0uis you point gluster at a mount
17:07 abradley it's the mount that I'm trying to create presently
17:10 Technicool joined #gluster
17:14 abradley http://paste.ubuntu.com/6193087/ os is installed on sda1. I want to use all of the remainder of the space on the disk as a brick. How do I move forward from here?
17:14 glusterbot Title: Ubuntu Pastebin (at paste.ubuntu.com)
17:15 abradley per the guide, mkfs.xfs -i size=512 /dev/sda2 returns "you must use -f"
17:15 abradley mkfs.xfs -f -i size=512 /dev/sda2 returns that paste
17:15 davinder joined #gluster
17:17 semiosis abradley: no one here can tell you which partition to format, you need to figure that out for yourself
17:17 abradley I'm not asking which partition to format
17:18 semiosis you could accidentally wipe out data & even wipe the computer clean (so it won't boot up anymore) if you do the wrong one
17:18 semiosis [13:05] <abradley> would sda2 be the correct choice?
17:18 abradley I'm asking "how do I make the command from the guide work?"
17:18 semiosis it works if you choose the right partition
17:18 abradley aHA! thanks
17:19 semiosis remember that the -f option for mkfx.xfs means "wipe clean any existing data on this partition" so BE CAREFUL
17:19 semiosis if you have a new partition with no data on it then you do not need -f
17:20 semiosis it would be safer to practice and learn this stuff inside a virtual machine
17:21 semiosis that way the most harm you can do is to a VM not your physical computer
17:21 semiosis take a look at oracle virtualbox -- it's very easy to use
17:21 semiosis i use that on my laptop for testing out glusterfs
17:24 abradley this is a vm
17:25 abradley http://paste.ubuntu.com/6193135/
17:25 glusterbot Title: Ubuntu Pastebin (at paste.ubuntu.com)
17:25 abradley I've tried every one of the options (besides the boot partition) and none of them work.    :/
17:26 abradley --every one of the listed partitions
17:26 semiosis hmm ok please pastie output of fdisk -l
17:26 l0uis abradley: sounds like you didn't make the partitions properly
17:26 semiosis oh never mind i see it in a previous paste
17:27 abradley http://paste.ubuntu.com/6193143/
17:27 glusterbot Title: Ubuntu Pastebin (at paste.ubuntu.com)
17:27 semiosis abradley: you have a vm with an 80 gig disk image?
17:27 abradley from a template, yes
17:27 semiosis pastie output of these commands please "pvdisplay; vgdisplay; lvdisplay"
17:28 bennyturns if I set a quota on /, should all directories and subdirectories inherit that quota or should it be for / only?
17:29 shylesh joined #gluster
17:30 abradley http://paste.ubuntu.com/6193151/
17:30 glusterbot Title: Ubuntu Pastebin (at paste.ubuntu.com)
17:30 l0uis abradley: if you're using lvm you should be making the xfs partition on the lvm device
17:30 l0uis s/partition/filesystem
17:30 abradley I see. I'll try and do that. thanks for the input
17:41 abradley I JUST created sda2 and am attempting to format it: http://paste.ubuntu.com/6193190/
17:41 glusterbot Title: Ubuntu Pastebin (at paste.ubuntu.com)
17:55 phox so this is a really fun one
17:55 phox gluster is being retarded and is giving me what I presume is -ENOENT for files that find is finding
17:55 phox if I MANUALLY go poke one of them it starts working thereafter
17:56 phox dunno if the problem ever comes back
17:56 phox but right now I have like 15-ish files in this dir that don't open using find ... | xargs md5sum
17:56 phox even though they exist.
17:56 phox so this sounds like a race, given that it seems to go away when I poke a SINGLE file manually
17:56 phox or maybe it has to do with find having just hit some level of cache or something
17:57 phox regardless, gluster is being retarded and I'd think this might be an interesting case of broken to examine :l
17:57 B219561 joined #gluster
18:05 failshell im kinda flabergasted. im trying to test a gluster setup in vagrant. i just cant peer probe my 2 vms
18:05 failshell just wont work
18:05 * failshell scratches his head
18:05 * failshell scratches his beard
18:08 semiosis failshell: what kinda network do your boxes have?
18:08 semiosis nat is :(
18:08 semiosis bridged is :)
18:08 failshell config.vm.network :private_network, ip: options["ipaddress"]
18:08 semiosis abradley: use lvm instead of raw partitions, since you have that set up
18:08 failshell i can ping both machines. i can telnet into the remote's machine ports
18:09 semiosis abradley: read the lvcreate man page... you can create an LV and then use that
18:09 abradley I'll give it a try
18:09 abradley thanks
18:10 failshell tcpdump time i guess
18:11 failshell semiosis: well, there's no traffic going from one to the other
18:11 semiosis failshell: iptables?
18:11 failshell nope
18:13 semiosis phox: check client log file... if a file is split brained the gluster client won't let you open it
18:13 semiosis other issues possible, but less likely.  log will say
18:20 failshell http://i.imgur.com/IgU4Y.gif
18:20 andreask joined #gluster
18:34 phox semiosis: one brick
18:34 * phox is always a source of breaking shit in new and exciting ways
18:39 failshell joined #gluster
18:39 davinder joined #gluster
18:46 ctria joined #gluster
18:52 failshell what port does gluster use during a peer probe?
18:52 failshell tcp/24007?
18:54 failshell 0-management: Peer uuid dc3d8170-bf9a-480c-92d0-9b8ccaf0f7bf is same as local uuid. Please check the uuid of both the peers from /var/lib/glusterd/glusterd.info
18:55 Remco @ports
18:55 glusterbot Remco: glusterd's management port is 24007/tcp and 24008/tcp if you use rdma. Bricks (glusterfsd) use 24009 & up for <3.4 and 49152 & up for 3.4. (Deleted volumes do not reset this counter.) Additionally it will listen on 38465-38467/tcp for nfs, also 38468 for NLM since 3.3.0. NFS also depends on rpcbind/portmap on port 111 and 2049 since 3.4.
18:58 johnmark JoeJulian: you around?
18:59 abradley I JUST created sda3. Why is CLI telling me "/dev/sda3: No such file or directory"?
18:59 abradley http://paste.ubuntu.com/6193517/
18:59 glusterbot Title: Ubuntu Pastebin (at paste.ubuntu.com)
19:12 wushudoin joined #gluster
19:17 zerick joined #gluster
19:21 rwheeler joined #gluster
19:23 abradley I just created a new partition /dev/sda4 in fdisk, accepting all defaults except the partition number. Why is the partition so small, shouldn't it, by default, use all the avaialbe space? http://paste.ubuntu.com/6193609/
19:23 glusterbot Title: Ubuntu Pastebin (at paste.ubuntu.com)
19:31 XpineX abradley: It seems you have created a primary partition, but almost all space on the drive has allready been allocated by an extended partition
19:38 semiosis abradley: i advised you to use LVM, why don't you do that?
19:46 abradley I'm not sure how to do that using cfdisk
19:46 abradley I JUST created sda2 with fdisk. Then I try to "mkfs.xfs...." and I get errors: http://paste.ubuntu.com/6193699/
19:46 glusterbot Title: Ubuntu Pastebin (at paste.ubuntu.com)
19:50 l0uis abradley: there error is right there, you didn't create a partition table.
19:50 abradley can cfdisk create a partition table?
19:51 abradley I don't understand. There are partitions there, yes? sda1 and sda2. Is it possible that there are partitions there without a partition table?
19:52 abradley doesn't fdisk -l display the partition table?
19:54 B219561 left #gluster
20:02 B21956 joined #gluster
20:02 B21956 left #gluster
20:18 semiosis abradley: your template VM uses LVM
20:18 semiosis so you should use LVM
20:18 semiosis or find another VM
20:18 semiosis s/find/find or build/
20:18 semiosis glusterbot: meh
20:18 glusterbot What semiosis meant to say was: use the latest GA release of glusterfs (currently 3.4.1) and report bugs if you find or build any
20:18 glusterbot semiosis: I'm not happy about it either
20:23 semiosis i'm sure this is confusing... VM = virtual machine; LVM = logical volume manager
20:25 Debolaz And not a highly intuitive one at that.
20:25 Debolaz :)
20:28 dtyarnell_ joined #gluster
20:32 adamb joined #gluster
20:48 Technicool joined #gluster
20:53 phox LVM = stupid
20:53 phox much simpler
21:49 Debolaz At times, I've contemplated writing a cli frontend to LVM to sidestep having to re-learn the utilities every time I want to use it.
21:49 failshel_ joined #gluster
22:04 \_pol joined #gluster
22:09 semiosis Debolaz: thats how NIH syndrome is born
22:11 * Debolaz thinks the LVM utilities suffers from RHD syndrome. (Red Hat Design)
22:12 semiosis haha
22:12 Debolaz Functional, fundamentally sound, but oh so ugly frontends to deal with.
22:13 Debolaz And that is true for almost everything I see coming out of Red Hat. :)
22:13 phox semiosis: so I looked into those -ENOENT (or whatever) files a bit.  it appears that there's nothing (about anything, at all) in the client log =/
22:13 phox nothing in the brick log either
22:13 semiosis phox: then i would try mounting a new client, see if it has the same problem... if it does, then i'd suggest stopping & starting the volume
22:14 phox happens on all clients
22:14 phox stop/start without client unmount?
22:14 semiosis client will stay mounted but won't work until the volume is started again iirc
22:14 * phox does that
22:14 phox yes
22:15 phox was wondering if you thought I should do a clean bringing up of the whole stack
22:15 semiosis i've seen something like that happen when i mixed incompatible versions of clients & servers
22:15 phox issue persists
22:15 phox 3.4.1 client + server
22:15 semiosis very weird
22:15 phox dunno if I can actually confirm that the FUSE crap was unloaded
22:15 phox this is post-upgrade and I did not reboot the clients but I unmounted _everything_ and re-mounted after all of the software upgrades
22:16 * phox gurgles ponderously
22:16 semiosis afk
22:55 nueces joined #gluster
23:07 srsc joined #gluster
23:08 srsc left #gluster
23:09 shane___ joined #gluster
23:55 zaitcev joined #gluster

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