#ceph IRC Log


IRC Log for 2015-04-01

Timestamps are in GMT/BST.

[14:39] -reticulum.oftc.net- *** Looking up your hostname...
[14:39] -reticulum.oftc.net- *** Checking Ident
[14:39] -reticulum.oftc.net- *** Couldn't look up your hostname
[14:39] -reticulum.oftc.net- *** No Ident response
[14:39] * CephLogBot (~PircBot@ has joined #ceph
[14:39] * Topic is 'CDS Schedule Posted: http://goo.gl/i72wN8 || http://ceph.com/get || dev channel #ceph-devel || test lab channel #sepia'
[14:39] * Set by scuttlemonkey!~scuttle@nat-pool-rdu-t.redhat.com on Mon Mar 02 21:13:33 CET 2015
[14:43] * maxxware (~maxx@ has joined #ceph
[14:44] * maxxware (~maxx@ has left #ceph
[14:44] * bene (~ben@nat-pool-bos-t.redhat.com) has joined #ceph
[14:44] <sugoruyo> is there a command to list rgw users?
[14:44] * jdillaman (~jdillaman@pool-173-66-110-250.washdc.fios.verizon.net) has joined #ceph
[14:46] * diegows (~diegows@ has joined #ceph
[14:46] * kefu (~kefu@ Quit (Max SendQ exceeded)
[14:47] * kefu (~kefu@ has joined #ceph
[14:47] * lalatenduM (~lalatendu@ Quit (Quit: Leaving)
[14:50] * AGaW (~Kurimus@5NZAAAZKU.tor-irc.dnsbl.oftc.net) Quit ()
[14:50] * visored (~Nephyrin@this-is-a-tor-exit-node-103.hviv.nl) has joined #ceph
[14:50] * lalatenduM (~lalatendu@ has joined #ceph
[14:53] * georgem (~Adium@fwnat.oicr.on.ca) has joined #ceph
[14:55] * diegows (~diegows@ Quit (Ping timeout: 480 seconds)
[14:59] * lpabon (~quassel@24-151-54-34.dhcp.nwtn.ct.charter.com) has joined #ceph
[15:06] * Hemanth (~Hemanth@ Quit (Ping timeout: 480 seconds)
[15:06] * amote (~amote@ Quit (Quit: Leaving)
[15:11] * xoritor (~xoritor@cpe-72-177-85-116.austin.res.rr.com) has joined #ceph
[15:13] * vbellur (~vijay@ Quit (Ping timeout: 480 seconds)
[15:14] * kanagaraj (~kanagaraj@ has joined #ceph
[15:16] * tupper (~tcole@2001:420:2280:1272:8900:f9b8:3b49:567e) has joined #ceph
[15:19] * cfreak200 (andi@p4FF3E9B0.dip0.t-ipconnect.de) has joined #ceph
[15:20] * visored (~Nephyrin@2WVAAARNX.tor-irc.dnsbl.oftc.net) Quit ()
[15:20] * Nephyrin (~Coe|work@edwardsnowden0.torservers.net) has joined #ceph
[15:28] * giorgis (~oftc-webi@ Quit (Quit: Page closed)
[15:30] * ircolle (~Adium@c-71-229-136-109.hsd1.co.comcast.net) has joined #ceph
[15:31] * haigang (~haigang@ has joined #ceph
[15:31] * vbellur (~vijay@ has joined #ceph
[15:36] * fam_away is now known as fam
[15:37] * arbrandes (~arbrandes@ has joined #ceph
[15:38] <cetex> how can i make osd's id dynamic?
[15:38] <cetex> -i $hostname.$disk?
[15:41] * danieagle (~Daniel@ has joined #ceph
[15:43] * xcezzz (~Adium@pool-100-3-14-19.tampfl.fios.verizon.net) has joined #ceph
[15:46] * diegows (~diegows@ has joined #ceph
[15:46] * rdas (~rdas@ Quit (Quit: Leaving)
[15:47] <bdonnahue> hello i am trying to start an osd (it shows in the osd tree) with ceph start osd.0 (on centos)
[15:47] <bdonnahue> the command fails though saying "osd.0 not found"
[15:47] <bdonnahue> can anyone help me out
[15:49] <championofcyrodi> bdonnahue: where are your osd logs writing to?
[15:49] <bdonnahue> /var/log/ceph/ceph-osd.0.log
[15:49] * rotbeard (~redbeard@b2b-94-79-138-170.unitymedia.biz) Quit (Quit: Leaving)
[15:50] * Nephyrin (~Coe|work@2WVAAARQE.tor-irc.dnsbl.oftc.net) Quit ()
[15:50] <visbits> rbd block device wont respond after sitting for a few hrs
[15:50] <visbits> have to remap them.. that seems to be the issue
[15:50] * Sami345 (~mollstam@ has joined #ceph
[15:50] <bdonnahue> looking at the logs. there is nothing in there (stale log left over from previous osd installation)
[15:50] * Nacer (~Nacer@252-87-190-213.intermediasud.com) Quit (Remote host closed the connection)
[15:51] <bdonnahue> maybe i need to purge and try again
[15:51] <championofcyrodi> what does 'ceph osd ls' display?
[15:52] <championofcyrodi> I have 7 osds, so for example, mine is showing 0-6
[15:52] <cetex> hm.
[15:52] * haigang (~haigang@ Quit (Quit: This computer has gone to sleep)
[15:52] <bdonnahue> it shows 0
[15:52] <cetex> or.. what are the constraints for osd id?
[15:52] <bdonnahue> i only have one osd named osd.0
[15:52] <cetex> can i have the same id on multiple nodes?
[15:52] <bdonnahue> cetex no
[15:52] <cetex> ok. so it needs to be unique in the dc
[15:53] * brutuscat (~brutuscat@93.Red-88-1-121.dynamicIP.rima-tde.net) has joined #ceph
[15:53] <bdonnahue> yes
[15:53] <cetex> if i want to have it deterministic and still automated i'd have to do something like $RACK$HOST$DISK then?
[15:53] <cetex> RACK = 01, HOST = 020, DISK = 004?
[15:53] <cetex> 01020004?
[15:54] <bdonnahue> whats your end game?
[15:54] <bdonnahue> goal*
[15:54] * kanagaraj (~kanagaraj@ Quit (Quit: Leaving)
[15:54] <cetex> large datacenter, all machines running the same pxe booted image, they need to determine disk-id automagically.
[15:54] <m0zes> for a while (not sure if this is still the case) ceph really did't like sparsely populated lists of osds.
[15:54] <bdonnahue> cetex the install process is capable of auto generating an id
[15:54] <visbits> v3493131: 10832 pgs: 9171 active+clean, 3 active+clean+scrubbing, 21 active+clean+scrubbing+deep, 38 active+remapped+backfilling, 1599 active+remapped+wait_backfill; 35444 GB data, 71644 GB used, 104 TB / 174 TB avail; -48444/20940729 objects degraded (-0.231%); 5263819/20940729 objects misplaced (25.137%); 599 MB/s, 155 objects/s recovering
[15:54] <cetex> so i guess i either need to generate it, or query ceph mon and then get the next id?
[15:54] <visbits> how can i have negitive objects degraded
[15:55] * dgurtner (~dgurtner@ Quit (Ping timeout: 480 seconds)
[15:55] * kefu (~kefu@ Quit (Max SendQ exceeded)
[15:55] <cetex> bdonnahue: yeah. but i'm not doing that. i need full control and i need to understand the setup in detail.
[15:55] <bdonnahue> I would not use the osd id to associate machine metadata
[15:55] <bdonnahue> instead I would use the ip
[15:56] <bdonnahue> and or hostname
[15:56] * kefu (~kefu@ has joined #ceph
[15:56] * wicope (~wicope@0001fd8a.user.oftc.net) has joined #ceph
[15:56] <bdonnahue> nothing outside the internal ceph cluster will care about the id
[15:56] <cetex> yeah. but how do i generate the id then :)
[15:56] <cetex> would be nice if i could just set it to "auto" if it's just a unique identifier.. :)
[15:57] <bdonnahue> cetex the default behavior is to generate an id when the osd is "created"
[15:57] <bdonnahue> if you specify no id then the next available id is chosen
[15:57] <cetex> aah. so i'm overengineering then. just remove id and it should work?
[15:57] <bdonnahue> yes
[15:59] <cetex> great.
[15:59] <cetex> how can i auto-add an osd to a host / rack / something so it's automatically in a "failure-domain" or something?
[15:59] <bdonnahue> cetex I would use ip's addresses to dictate the location of things
[16:00] <bdonnahue> many cloud provisioning tools like cloudstack for example understand this type of thing
[16:00] <bdonnahue> eg 192.datacenter.rack.host
[16:00] * hellertime (~Adium@a23-79-238-10.deploy.static.akamaitechnologies.com) Quit (Read error: Connection reset by peer)
[16:00] * hellertime (~Adium@ has joined #ceph
[16:00] <cetex> yeah.. we have highly specific requirements so we need to design that ourselves.
[16:01] <cetex> otherwise we end up with multiple cloud-management tools and a pile of shit :>
[16:01] * harold (~hamiller@71-94-227-66.dhcp.mdfd.or.charter.com) has joined #ceph
[16:01] <championofcyrodi> bdonnahue: I'm looking at the /etc/init.d/ceph script, because there is a hunk of code in there that determines osd ids validity
[16:02] * bene2 (~ben@nat-pool-bos-t.redhat.com) has joined #ceph
[16:02] <championofcyrodi> specifically this: http://paste.openstack.org/show/197781/
[16:02] <Be-El> cetex: ceph can read the location information from a script during osd startup
[16:02] <championofcyrodi> perhaps the ceph crush location hook
[16:02] * harold (~hamiller@71-94-227-66.dhcp.mdfd.or.charter.com) Quit ()
[16:03] <Be-El> cetex: http://ceph.com/docs/master/rados/operations/crush-map/
[16:03] <bdonnahue> championofcyrodi: giving it a look
[16:03] <Be-El> cetex: as championofcyrodi already mentioned, use the hool
[16:03] <Be-El> hook
[16:04] <cetex> hm, yeah. i'll do something like that when i get there. :)
[16:04] <cetex> currently it seems like -i needs to be specified to the osd
[16:04] <cetex> refuses to start without it.
[16:05] <Be-El> cetex: if you want to use the same image for the osds and are able to determine the correct crush location from a script, you don't need to care for osd ids at all
[16:06] <Be-El> cetex: osds are usually started based on the partition uuid of the osd data partition via a udev rule
[16:07] * cok (~chk@2a02:2350:18:1010:81c2:ccf6:dcb3:7589) Quit (Quit: Leaving.)
[16:07] <cetex> something like that is what i need to do.
[16:07] <cetex> 2015-04-01 14:03:14.643458 7f3cb0157800 0 ceph version 0.80.7 (6c0127fcb58008793d3c8b62d925bc91963672a3), process ceph-osd, pid 5
[16:07] <cetex> 2015-04-01 14:03:14.643468 7f3cb0157800 -1 must specify '-i #' where # is the osd number
[16:07] <cetex> that's what i get when i don't have the id set..
[16:07] * Nacer (~Nacer@252-87-190-213.intermediasud.com) has joined #ceph
[16:08] * bene (~ben@nat-pool-bos-t.redhat.com) Quit (Ping timeout: 480 seconds)
[16:09] <Be-El> are you trying to start the osd manually or with a custom init script?
[16:10] * b0e (~aledermue@ Quit (Quit: Leaving.)
[16:11] * nljmo (~nljmo@5ED6C263.cm-7-7d.dynamic.ziggo.nl) Quit (Quit: Textual IRC Client: www.textualapp.com)
[16:12] <cetex> currently with a custom init script.
[16:12] <cetex> which runs a docker container that mounts one of the harddrives as /data/ceph and then runs the osd.
[16:13] * yanzheng (~zhyan@ Quit (Quit: This computer has gone to sleep)
[16:14] <bdonnahue> i cannot get my osd to start. I keep seeing "osd.0- not found"
[16:14] <Be-El> do you prepare the harddrives with ceph-deploy or ceph-disk?
[16:14] <bdonnahue> *"osd.0 not found"
[16:15] <bdonnahue> i manually deployed this one
[16:15] <bdonnahue> i've been having problems with the ceph-deploy activate step of the process
[16:15] <Be-El> bdonnahue: is the osd listed in the output of 'ceph-disk list' ?
[16:16] <cetex> this is kinda what i'm after: if [ -f /data/ceph/fsid ]; then /usr/bin/ceph-disk -d prepare --data-dir /data/ceph/; ceph-osd -d --osd-data=/data/ceph/ --osd-journal=/data/ceph/journal --mkfs --mkjournal; else ceph-osd -d --osd-data=/data/ceph/ --osd-journal=/data/ceph/journal; fi'
[16:16] <cetex> if disk is setup for osd, run osd, otherwise set it up for osd
[16:16] <cetex> and then run it
[16:18] * hasues (~hasues@kwfw01.scrippsnetworksinteractive.com) has joined #ceph
[16:18] <cetex> i see i forgot an ! in the "if [ -f ..."
[16:18] <Be-El> cetex: /data/ceph/whoami should contain the osd id after successful osd creation
[16:19] <bdonnahue> http://pastebin.com/PXwDw8La
[16:19] * nljmo (~nljmo@5ED6C263.cm-7-7d.dynamic.ziggo.nl) has joined #ceph
[16:20] * Sami345 (~mollstam@425AAAFC0.tor-irc.dnsbl.oftc.net) Quit ()
[16:20] <Be-El> cetex: and you can probably skip the --osd-journal parameter upon starting the osd. the journal is excepted to be in that location by default
[16:20] <bdonnahue> Be-El thats the output of ceph-disk
[16:20] <visbits> Requires: librados2 = 1:0.80.7
[16:20] <visbits> Removing: 1:librados2-0.87-0.el7.centos.x86_64 (@Ceph)
[16:20] <visbits> librados2 = 1:0.87-0.el7.centos
[16:20] <visbits> Updated By: 1:librados2-0.87.1-0.el7.centos.x86_64 (Ceph)
[16:20] <visbits> librados2 = 1:0.87.1-0.el7.centos
[16:20] <visbits> Available: 1:librados2-0.86-0.el7.centos.x86_64 (Ceph)
[16:20] <visbits> librados2 = 1:0.86-0.el7.centos
[16:20] <visbits> so who fucked up the yum repo
[16:20] * hasues (~hasues@kwfw01.scrippsnetworksinteractive.com) has left #ceph
[16:21] <Be-El> bdonnahue: how do you start the osd? standard initscript/upstart? and which distribution do you use?
[16:21] <cetex> ah, great.
[16:21] * MACscr (~Adium@2601:d:c800:de3:f946:dcc5:5f15:a09e) Quit (Ping timeout: 480 seconds)
[16:21] <bdonnahue> im on centos. the docs list the init script as the start method
[16:22] * saltlake (~saltlake@ has joined #ceph
[16:22] <bdonnahue> i know that when I use ceph-deploy the os is started correctly
[16:22] <Be-El> bdonnahue: does the problem occur during manually starting via init script or during host boot up?
[16:24] * debian112 (~bcolbert@ has joined #ceph
[16:25] * dyasny (~dyasny@ Quit (Quit: Ex-Chat)
[16:25] * dyasny (~dyasny@ has joined #ceph
[16:26] <cetex> so.. if i do ceph-disk prepare ... and then ceph-osd --mkfs i should have the "whoami" file created (which i do)
[16:26] <cetex> whoami is the id?
[16:26] <cetex> why do i get a warning about the id later? :)
[16:28] <cetex> when i run "ceph-osd -d --osd-data=/data/ceph/"
[16:29] <Be-El> i think the ceph-disk prepare step includes creating the osd "filesystem", so you should be able to skip the -mkfs step
[16:29] <Be-El> ceph-disk is just a pythin script, so you can verify it in the source code
[16:30] <visbits> Be-El someone fucked the repo :(
[16:30] * joef (~Adium@2601:9:280:f2e:4c88:39e:9399:3911) has joined #ceph
[16:31] <cetex> ceph-disk does: ceph_fsid fsid magic
[16:31] <cetex> creates those
[16:32] <Be-El> visbits: you want to install a package that requires librados2 from firefly (0.80.7) and wants to remove librados2 from giant (0.87.7)
[16:33] * bigtoch (~bigtoch@ has joined #ceph
[16:33] <visbits> my repo are all set to giant, why would it target firefly
[16:33] <bigtoch> Hello
[16:33] <Be-El> cetex: ah, ok, i've hidden all the osd stuff in some puppet classes. has been some time since i had a look at them
[16:33] <bigtoch> Anybody alive on this channel?
[16:33] <Be-El> bigtoch: nope
[16:33] <cetex> :)
[16:33] <visbits> neg.
[16:33] <bigtoch> Hahaha
[16:33] <bigtoch> Hello Geeks
[16:33] * overclk (~overclk@ Quit (Quit: Leaving)
[16:34] <bigtoch> I need some help
[16:34] * bkopilov (~bkopilov@bzq-79-179-107-120.red.bezeqint.net) Quit (Ping timeout: 480 seconds)
[16:34] * bkopilov (~bkopilov@bzq-79-183-54-8.red.bezeqint.net) has joined #ceph
[16:34] <Be-El> visbits: but i'm using debian/ubuntu, so i'm afraid i cannot help you with that problem
[16:34] <bigtoch> I had an issue with the ceph rados gateway
[16:34] <bigtoch> I am able to create bucket using the s3 api
[16:34] <cetex> hm. no matter what it seems i need to specify the osdid on the commandline
[16:35] * dgurtner (~dgurtner@ has joined #ceph
[16:35] <magicrobotmonkey> visbits: i'd double check that you dont have both repos in yum.repos.d
[16:35] <cetex> and the osdid needs to be 1: globally unique, and i guess 2: quite static.
[16:35] <bigtoch> however whenever I want to create an object I got incomplete headers error message
[16:35] <visbits> yeah i just looked, all set to giant
[16:35] <bigtoch> Any idea?
[16:36] <visbits> your gonna love this, cobblers repo is conflicting but contains no ceph.
[16:36] <bdonnahue> ok i used ceph-deploy prepare and the osd was started ok
[16:37] <bdonnahue> i ust have missed a step or something with the manual deploy
[16:37] <cetex> so.. the "whoami", is that the osd id? or is it something else?
[16:37] <cetex> :>
[16:37] <bdonnahue> the problem im having now is that the activate statement fails
[16:37] <magicrobotmonkey> visbits: is it possible that firefly is coming from your repo or something?
[16:37] <magicrobotmonkey> can you install with --enablerepo
[16:38] <visbits> no my cobbler repo is just a centos base image
[16:38] <visbits> priority have it obsoleted
[16:38] <bdonnahue> http://pastebin.com/nw04Hky9
[16:39] <bdonnahue> does anyone have an idea of what is going wrong here?
[16:39] <visbits> http://ceph.com/rpm-giant/el7/x86_64/ its picking the 86 versions here
[16:39] <Be-El> cetex: it's the osd number that you need to pass as osd id
[16:41] <bigtoch> I have done some search but I could not get any answer
[16:41] <bigtoch> Can someone assist me?
[16:41] <Be-El> bdonnahue: ceph-disk (and thus ceph-deploy) insists on creating a GPT partition table in the device. this allows ceph-disk to set a partition uuid for the osd partition, which in turn is used by a udev rule to start the osd
[16:42] <xcezzz> isnt GPT also needed to properly partition disks larger than 2tb?
[16:43] <visbits> --> Running transaction check
[16:43] <visbits> ---> Package python-ceph.x86_64 1:0.87.1-0.el7.centos will be obsoleted
[16:43] <visbits> ---> Package python-ceph-compat.x86_64 1:0.80.7-0.4.el7 will be obsoleting
[16:43] <visbits> these are latest
[16:43] <visbits> doesnt make much sense
[16:44] <magicrobotmonkey> yea my yum is weak, so i cant help you much
[16:45] <off_rhoden> visbits: you want to 'echo "check_obsoletes=1" >> /etc/yum/pluginconf.d/priorities.conf'
[16:45] <off_rhoden> see: http://tracker.ceph.com/issues/10476
[16:45] <visbits> yep, have that
[16:45] * sjmtest (uid32746@id-32746.uxbridge.irccloud.com) Quit (Quit: Connection closed for inactivity)
[16:45] <off_rhoden> oh yeah? hmm
[16:45] <visbits> thats step 1 in my documentation
[16:45] <off_rhoden> :) well that's interesting...
[16:46] <bigtoch> Hello
[16:46] <bdonnahue> Be-El: whats confusing me is the error message about trying to mount the disk
[16:46] <visbits> off_rhoden http://pastebin.com/c7xaQWp5
[16:46] <bdonnahue> after running osd prepare my ceph status shows that the osd is up and in
[16:46] <Be-El> bdonnahue: trying to mount, trying to partition etc.
[16:46] * diegows (~diegows@ Quit (Ping timeout: 480 seconds)
[16:46] <bdonnahue> Be-El: what?
[16:47] <Be-El> bdonnahue: the part about /sbin/partx being unable to update the in-kernel partition table of the device
[16:47] <cetex> Be-El: yeah.. if ceph-osd during mkfs or something figures out the next osd id. how can i find that and feed it back to osd? :)
[16:48] * daniel2_ (~daniel2_@cpe-24-28-6-151.austin.res.rr.com) has joined #ceph
[16:48] <bdonnahue> cetex if you look at the docs about manually adding an osd, one of the commands outputs the osd id
[16:48] <bdonnahue> im not sure about using the auto scripts
[16:49] <bdonnahue> Be-El: im not sure why activate tries to mount the disk when prepare already renders the osd up and in
[16:49] <visbits> off_rhoden epel is broken
[16:49] * xcezzz (~Adium@pool-100-3-14-19.tampfl.fios.verizon.net) Quit (Read error: Connection reset by peer)
[16:49] <bdonnahue> for it to be up and in wouldnt it already have mounted and prepared the disk for use?
[16:49] * xcezzz (~Adium@pool-100-3-14-19.tampfl.fios.verizon.net) has joined #ceph
[16:49] <off_rhoden> visbits: I was about to look if there was an update on EPEL recently
[16:49] <Be-El> bdonnahue: maybe there's already a ceph-osd process running on the box?
[16:50] <bdonnahue> Be-El: there is. the osd.prepare statement starts the osd service
[16:50] <georgem> can somebody please point me to the S3 API code used by Ceph radosgw? where is it kept on github?
[16:50] <Be-El> bdonnahue: so the osd is up and running
[16:50] <off_rhoden> ktdreyer: I may need your help here to look at whether or not a new EPEL package is messing with us again. :)
[16:50] <bdonnahue> Be-El: yes
[16:50] <Be-El> bdonnahue: mission accomplished! \o/
[16:50] <visbits> off_rhoden disabled epel repo and was able to update everything without problem.
[16:50] <bdonnahue> mission not accomplished
[16:50] <cetex> ah.. right right
[16:50] <visbits> now latest giant (Y)
[16:50] <cetex> starting to make sense.
[16:51] <bdonnahue> it was put up and in without running the activate command
[16:51] <bdonnahue> are the docs out of date or am i misreading them?
[16:51] <bdonnahue> i thought the order of commands was "osd prepare" then "osd activate"
[16:51] <Be-El> bdonnahue: you only need to run the activate command manually if you do not use the partition uuid "magic"
[16:51] <off_rhoden> visbits: yeah, there is definitely some conflict there. There are package splits that are in EPEL, but not ceph.com packages. but installing on CentOS requires some other packages from EPEL. oy
[16:51] * alfredodeza (~alfredode@ has left #ceph
[16:51] <bdonnahue> ahh ok i must have misread the docs then
[16:52] <Be-El> bdonnahue: or the docs lack this information
[16:52] <visbits> off_rhoden someone needs to spend there week figuring out a stable package release schedule lol
[16:52] <visbits> nothing like testing your testing
[16:52] <visbits> now if i could just figure out why my cluster doesnt like client io and why i have negative degraded objects?
[16:52] <bdonnahue> so now that the osd is up and in my ceph status shows that i have ~3TB available and free but that 192 pgs are suck inactive and stuck unclean
[16:53] <bdonnahue> why are they not active and clean
[16:53] <ktdreyer> off_rhoden: sure, what's up?
[16:54] <off_rhoden> ktdreyer: http://pastebin.com/c7xaQWp5 looks like an EPEL update may be obsoleting upstream packages again. visbits says he has the check_obseletes flag set already
[16:54] <cetex> so, order is: generate uuid -> <ID>=ceph osd create -> ceph-disk prepare -i <ID> -> ceph-osd --mkfs -i <ID> -> ceph-osd -i <ID>
[16:54] <off_rhoden> I am looking to see if a package was recently updated
[16:54] * TMM (~hp@sams-office-nat.tomtomgroup.com) Quit (Quit: Ex-Chat)
[16:55] <ktdreyer> :( I don't think it was
[16:55] * gregmark (~Adium@ has joined #ceph
[16:55] <off_rhoden> hmm, yeah python-ceph-compat hasn't changed in 3 months.
[16:55] <ktdreyer> I suspect the patch at https://bugzilla.redhat.com/show_bug.cgi?id=1193182 will fix visbits's issue
[16:55] <ktdreyer> I just haven't had time to test it out :(
[16:56] <ktdreyer> thankfully this split is finally done upstream in Hammer
[16:56] <off_rhoden> ktdreyer: yeah, none of the python- packages for Ceph in EPEL have changed. so this shouldn't have been a new thing. weird.
[16:57] * purpleidea (~james@ has joined #ceph
[16:58] <ktdreyer> I have no idea why yum-priorities-plugin isn't working in the scenario you're describing. but it certainly sounds like issue 11104
[16:58] <kraken> ktdreyer might be talking about http://tracker.ceph.com/issues/11104 [dependency problem with Ceph "Giant" upstream and EPEL]
[16:59] * wschulze (~wschulze@ has joined #ceph
[16:59] <cetex> ceph osd create seems to create an uuid as well
[17:00] <cetex> *as well as ceph-disk prepare
[17:00] <cetex> who decides what's true?
[17:00] <cetex> if i do ceph osd create and get an id and a uuid
[17:00] <cetex> and then ceph-disk prepare on the path of the harddrive which generates an uuid
[17:01] <cetex> and then runs the osd with the id i received from ceph osd create
[17:01] <ktdreyer> off_rhoden: tell you what, I should be able to apply this patch from BZ 1193182 today and get it into epel-testing at least
[17:01] <cetex> which uuid will win?
[17:01] * PerlStalker (~PerlStalk@ has joined #ceph
[17:01] <off_rhoden> ktdreyer: cool
[17:03] <loicd> bigtoch: the question about rados is better asked here, more knowledgeable people to give you an answer ;-) (I don't know enough)
[17:03] <bdonnahue> i restarted the osd and the status changed to 192 creating+incomplete
[17:04] * nljmo (~nljmo@5ED6C263.cm-7-7d.dynamic.ziggo.nl) Quit (Read error: Connection reset by peer)
[17:04] * dgurtner (~dgurtner@ Quit (Read error: Connection reset by peer)
[17:04] <bdonnahue> there is however still a ceph health warning that 192 pgs incomlete, stuck inactive, and stuck unclean
[17:04] <bigtoch> loicd where can ask the qustion?
[17:04] <bdonnahue> not sure how to understand this
[17:04] <georgem> loicd: do you know where the source code for ceph is located? I'm looking for the S3 API part
[17:04] * nljmo (~nljmo@5ED6C263.cm-7-7d.dynamic.ziggo.nl) has joined #ceph
[17:06] <visbits> if restarting all the osd daemons on a 100 node cluster fixes a lot of stuff.. does that mean its buggy lol
[17:06] * arbrandes (~arbrandes@ Quit (Ping timeout: 480 seconds)
[17:07] <loicd> georgem: https://github.com/ceph/ceph/tree/master/src/rgw
[17:08] <georgem> loicd: thank you, I don't know how I didn't find it :(
[17:08] <loicd> :-)
[17:09] <bigtoch> loicd any idea about how to get solution?
[17:09] <loicd> bigtoch: unfortunately I don't have the solution
[17:10] <bigtoch> loicd Tx
[17:12] <bigtoch> Hello Geeks has naybody faced this issue before: radosgw incomplete headers while uploading an object onto a bucket using s3 api?
[17:14] * analbeard (~shw@support.memset.com) Quit (Quit: Leaving.)
[17:14] * jks (~jks@ Quit (Read error: No route to host)
[17:14] * jks (~jks@ has joined #ceph
[17:14] * joef (~Adium@2601:9:280:f2e:4c88:39e:9399:3911) Quit (Quit: Leaving.)
[17:15] <cetex> and it seems i do need to specify --journal
[17:15] <cetex> its pointing to /var/lib/something
[17:15] <cetex> otherwise.
[17:15] <cetex> :)
[17:20] * daniel2_ (~daniel2_@cpe-24-28-6-151.austin.res.rr.com) Quit (Quit: Textual IRC Client: www.textualapp.com)
[17:20] * puvo (~W|ldCraze@171.ip-5-135-148.eu) has joined #ceph
[17:20] * daniel2_ (~daniel2_@cpe-24-28-6-151.austin.res.rr.com) has joined #ceph
[17:23] <bdonnahue> ugh i added memory to my osd and when i booted the osd will not start
[17:23] <bdonnahue> ceph is complaining that osd0 is not found
[17:23] <bdonnahue> previously it was up and in before the reboot
[17:24] * alram (~alram@ has joined #ceph
[17:24] * tk12 (~tk12@ has joined #ceph
[17:25] * ircolle (~Adium@c-71-229-136-109.hsd1.co.comcast.net) Quit (Quit: Leaving.)
[17:25] <bdonnahue> there is nothing in the ceph0 dir
[17:26] <cetex> disc not mounted?
[17:26] <cetex> *disk
[17:28] <ktdreyer> off_rhoden: I did an EPEL 7 build with the versioned Obsoletes, but it seems there's a problem with EPEL 7's PPC buildroot: https://lists.fedoraproject.org/pipermail/epel-devel/2015-April/011030.html
[17:28] <ktdreyer> off_rhoden: when that's solved, I should be able to get a build out and pushed to epel-testing
[17:28] <sage> jamespage: any update on teh tcmalloc issue with trusty?
[17:28] * lalatenduM (~lalatendu@ Quit (Quit: Leaving)
[17:28] <jamespage> sage, urgh - sorry - no been a bit bust
[17:28] <tk12> I'm seeing blank lines in /var/log/ceph/ceph-osd.0.log. Did I misconfigure some log verbosity level somewhere?
[17:28] <jamespage> busy
[17:28] <jamespage> rhather
[17:28] * moore (~moore@ has joined #ceph
[17:28] <bdonnahue> ok the disk was not mounted
[17:28] <jamespage> sage, well maybe bust as well
[17:29] <bdonnahue> now there are files but i still canont start the osd
[17:30] * madkiss (~madkiss@2001:6f8:12c3:f00f:7553:23f9:e07d:7726) Quit (Remote host closed the connection)
[17:30] * madkiss (~madkiss@2001:6f8:12c3:f00f:957:16e9:43c4:c299) has joined #ceph
[17:31] * joshd1 (~jdurgin@68-119-140-18.dhcp.ahvl.nc.charter.com) has joined #ceph
[17:31] * bkopilov (~bkopilov@bzq-79-183-54-8.red.bezeqint.net) Quit (Ping timeout: 480 seconds)
[17:31] <sage> jamespage: no worries, consider this a gentle reminder. the sandisk folks are hitting this
[17:33] <jamespage> sage, I feel reminded - let me look in detail
[17:33] <bdonnahue> i dont understand why the osd shows in the osd tree but when i try to start it ceph says its not found
[17:35] * reed (~reed@75-101-54-131.dsl.static.fusionbroadband.com) has joined #ceph
[17:35] <sage> thanks!
[17:36] * eedten1 (~eedten@ Quit (Ping timeout: 480 seconds)
[17:37] <bdonnahue> the osd has a weight and the tree structure looks ok
[17:38] <bdonnahue> th ceph log doesnt show anything with resepct to this osd though
[17:39] * vbellur (~vijay@ Quit (Ping timeout: 480 seconds)
[17:40] <jamespage> sage, https://bugs.launchpad.net/ubuntu/+source/google-perftools/+bug/1439277
[17:42] * fam is now known as fam_away
[17:44] * lpabon (~quassel@24-151-54-34.dhcp.nwtn.ct.charter.com) Quit (Remote host closed the connection)
[17:45] * brutuscat (~brutuscat@93.Red-88-1-121.dynamicIP.rima-tde.net) Quit (Remote host closed the connection)
[17:46] <bdonnahue> ok giving up purging trying out firefly
[17:47] <visbits> Be-El restarting all of my OSD fixed everything it appears..
[17:47] * dgurtner (~dgurtner@ has joined #ceph
[17:49] * dugravot6 (~dugravot6@dn-infra-04.lionnois.univ-lorraine.fr) Quit (Quit: Leaving.)
[17:50] * puvo (~W|ldCraze@5NZAAAZTN.tor-irc.dnsbl.oftc.net) Quit ()
[17:50] * FNugget (~Random@tor-exit.server9.tvdw.eu) has joined #ceph
[17:50] * arbrandes (~arbrandes@ has joined #ceph
[17:53] * purpleidea is now known as Guest730
[17:53] * purpleidea (~james@ has joined #ceph
[17:55] * Guest730 (~james@ Quit (Ping timeout: 480 seconds)
[17:56] <jamespage> sage, ok - that looks fine for SRU - just need to get a reproducer if possible for the SRU process
[17:57] <sage> ok. can you reply to somnath? he can probably help with that
[17:57] <sage> ah, you just did. :)
[17:59] * Rickus (~Rickus@office.protected.ca) Quit (Read error: No route to host)
[18:00] * puffy (~puffy@ has joined #ceph
[18:01] * Sysadmin88 (~IceChat77@ has joined #ceph
[18:02] * bandrus (~brian@207.sub-70-211-69.myvzw.com) has joined #ceph
[18:07] * kawa2014 (~kawa@ Quit (Quit: Leaving)
[18:08] <Vivek> I am getting a warning message as follows http://paste.ubuntu.com/10719059/
[18:09] * joshd1 (~jdurgin@68-119-140-18.dhcp.ahvl.nc.charter.com) Quit (Quit: Leaving.)
[18:09] <Vivek> What could be the reason ?
[18:09] <Vivek> Hi sage.
[18:10] <Vivek> loicd: Are you around and free ?
[18:10] * brutuscat (~brutuscat@93.Red-88-1-121.dynamicIP.rima-tde.net) has joined #ceph
[18:11] <loicd> Vivek: hi ! I'd be happy to answer a quick question before going home ;-)
[18:12] <Vivek> Can you tell me why I am getting that health warning
[18:13] <Vivek> I have 20 Osds and osd_pool_default_pg_num = 1024
[18:13] <Vivek> osd_pool_default_pgp_num = 1024
[18:13] <Vivek> I incresed it to 1024 from 512
[18:13] <Vivek> (20*1000)/3
[18:13] * xarses (~andreww@c-76-126-112-92.hsd1.ca.comcast.net) Quit (Ping timeout: 480 seconds)
[18:14] <Vivek> 1024 is nearest power of 2
[18:14] <Vivek> ceph -w still shows the same warning.
[18:14] <Vivek> s/incresed/increased
[18:17] * Kioob`Taff (~plug-oliv@2a01:e35:2e8a:1e0::42:10) Quit (Quit: Leaving.)
[18:17] <Vivek> loicd: What could be the reason ?
[18:18] <loicd> the warning is more likely because you have 64 stale+remapped+incomplete pgs
[18:20] * FNugget (~Random@98EAAAS1R.tor-irc.dnsbl.oftc.net) Quit ()
[18:20] * homosaur (~Linkshot@ has joined #ceph
[18:22] * sjm (~sjm@pool-173-70-76-86.nwrknj.fios.verizon.net) Quit (Read error: No route to host)
[18:23] * bigtoch (~bigtoch@ Quit (Read error: Connection reset by peer)
[18:23] * sjm (~sjm@pool-173-70-76-86.nwrknj.fios.verizon.net) has joined #ceph
[18:23] * puffy (~puffy@ Quit (Quit: Leaving.)
[18:23] * dupont-y (~dupont-y@2a01:e34:ec92:8070:353b:4f62:aa4c:134b) has joined #ceph
[18:27] <Vivek> loicd: ok
[18:27] <Vivek> loicd: How do I solve this issue ?
[18:27] * rongze (~rongze@ has joined #ceph
[18:28] <loicd> Vivek: following http://ceph.com/docs/master/rados/troubleshooting/troubleshooting-osd/ would be my first choice
[18:28] <loicd> sorry http://ceph.com/docs/master/rados/troubleshooting/troubleshooting-pg/
[18:29] <Vivek> Ok.
[18:29] * Rickus (~Rickus@office.protected.ca) has joined #ceph
[18:29] * _prime_ (~oftc-webi@ has joined #ceph
[18:34] * alram (~alram@ Quit (Quit: leaving)
[18:35] * fghaas (~florian@91-119-140-224.dynamic.xdsl-line.inode.at) Quit (Quit: Leaving.)
[18:36] * brutuscat (~brutuscat@93.Red-88-1-121.dynamicIP.rima-tde.net) Quit (Remote host closed the connection)
[18:39] * alram (~alram@ has joined #ceph
[18:42] * brutuscat (~brutuscat@93.Red-88-1-121.dynamicIP.rima-tde.net) has joined #ceph
[18:43] * wschulze (~wschulze@ Quit (Quit: Leaving.)
[18:49] * ChrisNBl_ (~ChrisNBlu@ Quit (Quit: My Mac has gone to sleep. ZZZzzz???)
[18:50] * homosaur (~Linkshot@2WVAAAR8Q.tor-irc.dnsbl.oftc.net) Quit ()
[18:51] * jordanP (~jordan@ Quit (Quit: Leaving)
[18:52] * gregmark (~Adium@ Quit (Quit: Leaving.)
[18:53] * xarses (~andreww@ has joined #ceph
[19:01] <georgem> Vivek: it says you have 65 pgs, did you increase both pg_num and pgp_num?
[19:02] * ChrisNBlum (~ChrisNBlu@ has joined #ceph
[19:03] <Vivek> Yes from 512 to 1024 in ceph.conf
[19:04] * xarses_ (~andreww@ has joined #ceph
[19:04] * cholcombe973 (~chris@7208-76ef-ff1f-ed2f-329a-f002-3420-2062.6rd.ip6.sonic.net) has joined #ceph
[19:04] * rongze (~rongze@ Quit (Ping timeout: 480 seconds)
[19:05] * fghaas (~florian@91-119-140-224.dynamic.xdsl-line.inode.at) has joined #ceph
[19:06] <Vivek> georgem: ping
[19:07] * wschulze (~wschulze@ has joined #ceph
[19:09] * xarses (~andreww@ Quit (Ping timeout: 480 seconds)
[19:11] * dgurtner (~dgurtner@ Quit (Ping timeout: 480 seconds)
[19:15] * wschulze (~wschulze@ Quit (Ping timeout: 480 seconds)
[19:15] * oro (~oro@2001:620:20:16:28ca:c4dd:3e62:ceeb) Quit (Ping timeout: 480 seconds)
[19:15] <theanalyst> Increasing in ceph.conf wont affect already created pools for which you've to follow https://ceph.com/docs/master/rados/operations/placement-groups/#set-the-number-of-placement-groups
[19:16] * wschulze (~wschulze@nat-pool-rdu-t.redhat.com) has joined #ceph
[19:17] * kefu (~kefu@ Quit (Quit: My Mac has gone to sleep. ZZZzzz???)
[19:20] * totalwormage (~Nephyrin@5NZAAAZ0B.tor-irc.dnsbl.oftc.net) has joined #ceph
[19:21] * Nacer (~Nacer@252-87-190-213.intermediasud.com) Quit (Ping timeout: 480 seconds)
[19:23] <Vivek> theanalyst: ok
[19:23] * mattronix (~quassel@fw1.sdc.mattronix.nl) Quit (Read error: Connection reset by peer)
[19:27] * brutuscat (~brutuscat@93.Red-88-1-121.dynamicIP.rima-tde.net) Quit (Remote host closed the connection)
[19:28] * mattronix (~quassel@fw1.sdc.mattronix.nl) has joined #ceph
[19:28] * brutuscat (~brutuscat@93.Red-88-1-121.dynamicIP.rima-tde.net) has joined #ceph
[19:29] * nils__ (~nils@doomstreet.collins.kg) Quit (Read error: Connection reset by peer)
[19:31] * alram (~alram@ Quit (Quit: leaving)
[19:33] * brutuscat (~brutuscat@93.Red-88-1-121.dynamicIP.rima-tde.net) Quit (Remote host closed the connection)
[19:33] * mykola (~Mikolaj@ has joined #ceph
[19:36] * diegows (~diegows@ has joined #ceph
[19:39] * vbellur (~vijay@ has joined #ceph
[19:39] * nils__ (~nils@doomstreet.collins.kg) has joined #ceph
[19:42] * lalatenduM (~lalatendu@ has joined #ceph
[19:44] * jwilkins (~jwilkins@2601:9:4580:f4c:ea2a:eaff:fe08:3f1d) has joined #ceph
[19:45] * puffy (~puffy@ has joined #ceph
[19:48] <bdonnahue> hello is anyone familiar with ceph fuse
[19:49] * nils______ (~nils@doomstreet.collins.kg) has joined #ceph
[19:50] * totalwormage (~Nephyrin@5NZAAAZ0B.tor-irc.dnsbl.oftc.net) Quit ()
[19:50] * zviratko (~Drezil@ has joined #ceph
[19:51] <bdonnahue> im wanted to test it out so i installed cephfuse on one of my osds
[19:52] <bdonnahue> i thn try to mount the cluster but the command seems to hang
[19:53] * Mika_c (~quassel@59-115-218-49.dynamic.hinet.net) has joined #ceph
[19:54] * Mika_c (~quassel@59-115-218-49.dynamic.hinet.net) Quit (Remote host closed the connection)
[19:54] * bitserker (~toni@63.pool85-52-240.static.orange.es) Quit (Ping timeout: 480 seconds)
[19:54] * nils__ (~nils@doomstreet.collins.kg) Quit (Ping timeout: 480 seconds)
[19:55] <cetex> wohoo!
[19:55] <cetex> mon's and osd's running :)
[19:55] * Nikos (~Nikos@85-127-206-247.dynamic.xdsl-line.inode.at) has joined #ceph
[19:56] <Nikos> hi ... i have a problem with a ceph client ... i use an rdb device, and mkfs.ext4 hangs ...
[19:56] <gregsfortytwo> bdonnahue: that probably means the MDS isn't running or something; you need to set up pools for the FS and run an MDS
[19:56] <gregsfortytwo> see http://ceph.com/docs/master/start/quick-cephfs/
[19:56] <gregsfortytwo> and http://ceph.com/docs/master/cephfs/ for more generic info
[19:57] * Concubidated (~Adium@ has joined #ceph
[19:58] * oro (~oro@80-219-254-208.dclient.hispeed.ch) has joined #ceph
[19:59] * purpleidea is now known as Guest749
[19:59] <bdonnahue> gregsfortytwo the cluster is u all monitors are in quorum 64pgs are active and clean
[19:59] * wschulze1 (~wschulze@ has joined #ceph
[19:59] * purpleidea (~james@ has joined #ceph
[19:59] <bdonnahue> i just performed a clean install
[20:00] <gregsfortytwo> yes, so you haven't set up pools for the FS or turned on an MDS
[20:00] <gregsfortytwo> there are extra steps to get a filesystem; refer to the documentation for more info :)
[20:00] * georege (~artemfn@2a02:6b8:0:408:f21f:afff:fe5f:aa6a) has joined #ceph
[20:00] <bdonnahue> AHHH im an idiot i read mds but by brain thought mon
[20:00] <bdonnahue> thanks
[20:00] * appaji (~appaji@ has joined #ceph
[20:01] <gregsfortytwo> :)
[20:01] <georege> Hello. I'm trying to create a cluster on a signle node by using this tutorial. http://dachary.org/?p=2374 I'm doing ceph osd crush set osd.0 1 root=default host=loclahost but ceph says me unknown command crush (-22) what can this be?
[20:01] * Guest749 (~james@ Quit (Ping timeout: 480 seconds)
[20:02] <appaji> Are there deployments of ceph out there using virtual disks in VMs as OSDs?
[20:02] * wschulze1 (~wschulze@ Quit ()
[20:03] <appaji> I know that the documentation strongly suggests that physical disks not be shared, even between two OSDs, let alone across applications.
[20:03] <bdonnahue> ok i created the mds but i dont see them listed with ceph status
[20:03] <bdonnahue> is there another command to show me themds servers
[20:05] <Nikos> ceph mds stat
[20:05] <Nikos> for example
[20:05] * ircolle (~Adium@c-71-229-136-109.hsd1.co.comcast.net) has joined #ceph
[20:06] * diegows (~diegows@ Quit (Ping timeout: 480 seconds)
[20:06] <bdonnahue> e1 000 up
[20:06] * wschulze (~wschulze@nat-pool-rdu-t.redhat.com) Quit (Ping timeout: 480 seconds)
[20:06] <bdonnahue> grr im rdp over https and i think my hosts function lock is diabled
[20:07] <bdonnahue> i think the mds servers are up i am still not able to mount with ceph fuse though
[20:07] * derjohn_mob (~aj@fw.gkh-setu.de) Quit (Ping timeout: 480 seconds)
[20:08] <bdonnahue> im reading the docs about creating a pool
[20:09] <bdonnahue> im slightly confused i thought the osds took care ofthat
[20:09] <Nikos> creating a pool is easy ... ceph osd pool create testpool 256 256 replicated
[20:10] <Nikos> with the 256 te amount of placement groups depends on your amount of osd's
[20:10] <bdonnahue> Nikos thanks im reading the docs now
[20:10] <bdonnahue> so what is the design behind placement groups
[20:10] <bdonnahue> currently i have 3 osds
[20:11] <georege> ceph osd crush set osd.0 1 root=default why this command says "Unknown command crush"?
[20:12] <georege> I've issued ceph osd create which created an osd with id 0
[20:12] <Nikos> bdonnahue: A placement group (PG) aggregates objects within a pool because tracking object placement and object metadata on a per-object basis is computationally expensive???i.e., a system with millions of objects cannot realistically track placement on a per-object basis. ( i copied that from the docs, because it is a good explanation )
[20:13] <bdonnahue> gregsfortytwo Nikos When i run ceph status it shows that there is already 1 pool
[20:13] <bdonnahue> thanks Nikos
[20:13] <bdonnahue> the docs ask me to create another pool for data and one for meta data
[20:14] * Hemanth (~Hemanth@ has joined #ceph
[20:14] <bdonnahue> im confused what this default pool is that i see now
[20:15] <Nikos> when you create a ceph cluster, the "rbd" pool is always created
[20:15] <bdonnahue> ahh ok that makes sense thanks
[20:16] <bdonnahue> is there a command to list the pools
[20:17] <Nikos> ofc ... ceph osd lspools
[20:17] <bdonnahue> got it ceph osd lspools
[20:17] <Nikos> ^^
[20:17] <bdonnahue> beat me to it thanks
[20:17] * purpleidea (~james@ Quit (Ping timeout: 480 seconds)
[20:17] <bdonnahue> so after creating the pools my status has changed from 64 activeclean
[20:19] <bdonnahue> ok now it is 192 active clean
[20:20] * zviratko (~Drezil@5NZAAAZ2V.tor-irc.dnsbl.oftc.net) Quit ()
[20:21] * purpleidea (~james@ has joined #ceph
[20:21] <Nikos> yes, well, there are more active pg's now
[20:23] * bkopilov (~bkopilov@bzq-79-178-196-42.red.bezeqint.net) has joined #ceph
[20:23] * Be-El (~quassel@fb08-bcf-pc01.computational.bio.uni-giessen.de) Quit (Remote host closed the connection)
[20:25] <Guest646> re: epel issues, i was able to do --exclude=*0.80.7* and get around it
[20:25] * Guest646 is now known as Loth
[20:27] * Nacer (~Nacer@203-206-190-109.dsl.ovh.fr) has joined #ceph
[20:27] <georege> ceph osd crush set osd.0 1 root=default host=
[20:27] <georege> Error ENOENT: unable to set item id 0 name 'osd.0' weight 1 at location {host=,root=default}: does not exist
[20:28] <georege> What could be the reason for this?
[20:31] <Nikos> what exactly are you trying to do ?
[20:33] <Nikos> i mean, since the osd has a distinct name anyway, ceph osd crush set osd.0 1 should suffice, i think ( not entirely sure, but ... )
[20:34] <georege> Okay, maybe I don't need to do this. I need to create a cluster in a single folder. http://lpaste.net/129982
[20:35] <georege> And probably rados command hangs because I've not set osd replication count?
[20:35] * bandrus1 (~brian@207.sub-70-211-69.myvzw.com) has joined #ceph
[20:36] <georege> which I'm trying to set with ceph osd crush set osd.0 1
[20:36] <georege> but it says http://lpaste.net/129983
[20:36] * ChrisNBlum (~ChrisNBlu@ Quit (Quit: My Mac has gone to sleep. ZZZzzz???)
[20:38] * georege (~artemfn@2a02:6b8:0:408:f21f:afff:fe5f:aa6a) Quit (Quit: WeeChat 1.0)
[20:38] * bandrus (~brian@207.sub-70-211-69.myvzw.com) Quit (Ping timeout: 480 seconds)
[20:39] * alram (~alram@ has joined #ceph
[20:39] * clayb (~clayb@ has joined #ceph
[20:41] * purpleidea is now known as Guest750
[20:42] * purpleidea (~james@ has joined #ceph
[20:42] * Hemanth (~Hemanth@ Quit (Ping timeout: 480 seconds)
[20:44] <Vivek> georgem: My issue is resolved, the rbd pool was deleted as it had not data and was re-created with pg and pgp_num as 1024.
[20:44] <Vivek> ceph health is now ok.
[20:44] <georgem> col
[20:44] <Vivek> d/not/no
[20:44] * Guest750 (~james@ Quit (Ping timeout: 480 seconds)
[20:45] <Vivek> s/not/no
[20:45] <Vivek> I am off to bed.
[20:45] <georgem> s/col/cool
[20:45] <Vivek> georgem: Thanks for your help.
[20:45] <Vivek> Bye.
[20:47] * jo00nas (~jonas@188-183-5-254-static.dk.customer.tdc.net) has joined #ceph
[20:48] * jo00nas (~jonas@188-183-5-254-static.dk.customer.tdc.net) Quit ()
[20:54] * clayb (~clayb@ Quit (Quit: Leaving.)
[20:54] * Drezil (~QuantumBe@balo.jager.io) has joined #ceph
[20:55] * B_Rake (~B_Rake@69-195-66-67.unifiedlayer.com) has joined #ceph
[20:56] * nhm_ (~nhm@184-97-225-22.mpls.qwest.net) has joined #ceph
[20:57] * MACscr (~Adium@2601:d:c800:de3:d943:265e:a5d6:a9c4) has joined #ceph
[21:00] * sbfox (~Adium@ has joined #ceph
[21:00] * bkopilov (~bkopilov@bzq-79-178-196-42.red.bezeqint.net) Quit (Ping timeout: 480 seconds)
[21:01] * alram (~alram@ Quit (Quit: leaving)
[21:02] * nhm (~nhm@65-128-142-103.mpls.qwest.net) Quit (Ping timeout: 480 seconds)
[21:03] * ngoswami (~ngoswami@ Quit (Quit: Leaving)
[21:09] * appaji (~appaji@ Quit (Quit: Adios amigos.)
[21:10] * rotbeard (~redbeard@2a02:908:df10:d300:76f0:6dff:fe3b:994d) has joined #ceph
[21:11] * vata (~vata@culture.mcc.gouv.qc.ca) has joined #ceph
[21:11] * branto (~branto@ip-213-220-214-203.net.upcbroadband.cz) has left #ceph
[21:17] * nhm_ (~nhm@184-97-225-22.mpls.qwest.net) Quit (Ping timeout: 480 seconds)
[21:20] * hellertime (~Adium@ Quit (Quit: Leaving.)
[21:20] * nhm (~nhm@184-97-245-198.mpls.qwest.net) has joined #ceph
[21:20] * ChanServ sets mode +o nhm
[21:24] * Drezil (~QuantumBe@5NZAAAZ6G.tor-irc.dnsbl.oftc.net) Quit ()
[21:24] * nicatronTg (~Zeis@tor-exit.server7.tvdw.eu) has joined #ceph
[21:28] * sjmtest (uid32746@id-32746.uxbridge.irccloud.com) has joined #ceph
[21:30] * dgurtner (~dgurtner@ has joined #ceph
[21:35] * purpleidea is now known as Guest754
[21:35] * purpleidea (~james@ has joined #ceph
[21:37] * Guest754 (~james@ Quit (Ping timeout: 480 seconds)
[21:45] * thomnico (~thomnico@2a01:e35:8b41:120:287f:5e1f:458e:1b9a) Quit (Ping timeout: 480 seconds)
[21:50] * lalatenduM (~lalatendu@ Quit (Quit: Leaving)
[21:52] * Nikos (~Nikos@0001ff95.user.oftc.net) has left #ceph
[21:54] * nicatronTg (~Zeis@5NZAAAZ7I.tor-irc.dnsbl.oftc.net) Quit ()
[21:54] * darkid (~Freddy@aurora.enn.lu) has joined #ceph
[21:57] * rendar (~I@host45-128-dynamic.61-82-r.retail.telecomitalia.it) Quit (Ping timeout: 480 seconds)
[21:59] * rendar (~I@host45-128-dynamic.61-82-r.retail.telecomitalia.it) has joined #ceph
[22:00] * georgem (~Adium@fwnat.oicr.on.ca) has left #ceph
[22:01] * t0rn (~ssullivan@2607:fad0:32:a02:d227:88ff:fe02:9896) Quit (Quit: Leaving.)
[22:02] * derjohn_mob (~aj@tmo-110-125.customers.d1-online.com) has joined #ceph
[22:04] * lcurtis (~lcurtis@ has joined #ceph
[22:04] * ganders (~root@ Quit (Quit: WeeChat 0.4.2)
[22:12] * derjohn_mob (~aj@tmo-110-125.customers.d1-online.com) Quit (Remote host closed the connection)
[22:12] * sherlocked (~watson@ has joined #ceph
[22:14] * vata (~vata@culture.mcc.gouv.qc.ca) Quit (Ping timeout: 480 seconds)
[22:22] * alfredodeza (~alfredode@ has joined #ceph
[22:24] * darkid (~Freddy@2WVAAASVK.tor-irc.dnsbl.oftc.net) Quit ()
[22:24] * Popz (~biGGer@ has joined #ceph
[22:28] * vata (~vata@ has joined #ceph
[22:31] * nils______ (~nils@doomstreet.collins.kg) Quit (Quit: This computer has gone to sleep)
[22:35] * nils______ (~nils@doomstreet.collins.kg) has joined #ceph
[22:37] * dgurtner (~dgurtner@ Quit (Ping timeout: 480 seconds)
[22:38] * nhm_ (~nhm@65-128-146-154.mpls.qwest.net) has joined #ceph
[22:38] * purpleidea (~james@ Quit (Ping timeout: 480 seconds)
[22:39] * dupont-y (~dupont-y@2a01:e34:ec92:8070:353b:4f62:aa4c:134b) Quit (Ping timeout: 480 seconds)
[22:41] * DV (~veillard@2001:41d0:a:f29f::1) Quit (Remote host closed the connection)
[22:41] * purpleidea (~james@ has joined #ceph
[22:42] * DV (~veillard@2001:41d0:1:d478::1) has joined #ceph
[22:44] * nhm (~nhm@184-97-245-198.mpls.qwest.net) Quit (Ping timeout: 480 seconds)
[22:47] * mliang2 (~oftc-webi@ has joined #ceph
[22:47] <mliang2> hi, anyone w/ experence on ceph's cache tier?
[22:53] * saltlake (~saltlake@ Quit (Ping timeout: 480 seconds)
[22:54] * Popz (~biGGer@2WVAAASXY.tor-irc.dnsbl.oftc.net) Quit ()
[22:54] * brianjjo (~rcfighter@hessel3.torservers.net) has joined #ceph
[22:55] * bkopilov (~bkopilov@bzq-79-183-28-54.red.bezeqint.net) has joined #ceph
[22:59] * doppelgrau (~doppelgra@pd956d116.dip0.t-ipconnect.de) has joined #ceph
[22:59] * tupper (~tcole@2001:420:2280:1272:8900:f9b8:3b49:567e) Quit (Ping timeout: 480 seconds)
[23:00] * sbfox (~Adium@ Quit (Quit: Leaving.)
[23:00] * eoin (~eoin@89-93-153-201.hfc.dyn.abo.bbox.fr) has joined #ceph
[23:03] * purpleidea (~james@ Quit (Ping timeout: 480 seconds)
[23:04] * bandrus1 (~brian@207.sub-70-211-69.myvzw.com) Quit (Ping timeout: 480 seconds)
[23:06] * i_m (~ivan.miro@nat-5-carp.hcn-strela.ru) Quit (Ping timeout: 480 seconds)
[23:06] * diegows (~diegows@ has joined #ceph
[23:07] * diegows (~diegows@ Quit ()
[23:09] * jdillaman (~jdillaman@pool-173-66-110-250.washdc.fios.verizon.net) Quit (Quit: jdillaman)
[23:12] * oro (~oro@80-219-254-208.dclient.hispeed.ch) Quit (Ping timeout: 480 seconds)
[23:12] * jdillaman (~jdillaman@pool-173-66-110-250.washdc.fios.verizon.net) has joined #ceph
[23:12] * wicope (~wicope@0001fd8a.user.oftc.net) Quit (Remote host closed the connection)
[23:13] * reed (~reed@75-101-54-131.dsl.static.fusionbroadband.com) Quit (Quit: Ex-Chat)
[23:14] * analbeard (~shw@support.memset.com) has joined #ceph
[23:14] <bdonnahue> hello. i was able to mount my ceph file system but i have a question. I mounted an ip that did not belong to the active mds
[23:15] * bandrus (~brian@207.sub-70-211-69.myvzw.com) has joined #ceph
[23:15] <bdonnahue> i guess mount goes through the monito since it uses port 6789?
[23:15] * reed (~reed@75-101-54-131.dsl.static.fusionbroadband.com) has joined #ceph
[23:18] * ChrisNBlum (~ChrisNBlu@ has joined #ceph
[23:21] <jcsp1> bdonnahue: the address you specify in mount is always your mon. If you have multiple mons, specify all of them separated by commas (http://ceph.com/docs/master/man/8/mount.ceph/)
[23:22] * cholcombe973 (~chris@7208-76ef-ff1f-ed2f-329a-f002-3420-2062.6rd.ip6.sonic.net) Quit (Quit: Leaving.)
[23:22] * clayb (~clayb@ has joined #ceph
[23:23] * clayb (~clayb@ Quit ()
[23:23] * analbeard (~shw@support.memset.com) has left #ceph
[23:24] * brianjjo (~rcfighter@2WVAAAS0B.tor-irc.dnsbl.oftc.net) Quit ()
[23:24] <bdonnahue> jcsp1: thanks for the info. I was thinking of putting a vip in front of the mons
[23:24] * sbfox (~Adium@ has joined #ceph
[23:24] * Deiz (~Jones@tor-exit.server7.tvdw.eu) has joined #ceph
[23:25] <bdonnahue> im not sure how this vip would detect that a mon is out of quorum
[23:25] <bdonnahue> any thoughts?
[23:25] * wschulze (~wschulze@cpe-74-73-11-233.nyc.res.rr.com) has joined #ceph
[23:25] <dmick> why would you do that?
[23:25] <bilco105> bdonnahue: the IP address needs to be unique to the mon, therefore a VIP wouldn't work anyway
[23:27] <lcurtis> anyone using rgw here?
[23:27] * reed (~reed@75-101-54-131.dsl.static.fusionbroadband.com) Quit (Quit: Ex-Chat)
[23:28] * joef1 (~Adium@2601:9:280:f2e:7529:bdc6:30f8:1a84) has joined #ceph
[23:28] * joef1 (~Adium@2601:9:280:f2e:7529:bdc6:30f8:1a84) has left #ceph
[23:29] <bdonnahue> bilco105: dmick. I am trying to export my cephfs over nfs and use it as a datastore for various things that dont have access to the ceph fue driver
[23:29] <bdonnahue> why wouldnt a vip work?
[23:29] <bdonnahue> if mons are .1, .2, and .3 the vip could be .0
[23:30] <dmick> the ceph client needs to know about the mons. that's how it works.
[23:30] <dmick> exporting the cephfs over nfs isn't affected by that
[23:30] <bdonnahue> ahh ok so then i just comma delimit the list of mons
[23:30] <bilco105> bdonnahue: either that or a single mon IP
[23:30] <m0zes> how about rrdns for the mon name resolution?
[23:31] <bilco105> ceph will learn the rest
[23:31] * _prime_ (~oftc-webi@ Quit (Quit: Page closed)
[23:31] <bdonnahue> bilco105 thanks. One other question I had was how can i tell ceph which nic to listen on? i want it to go through a vip rather than
[23:31] <bilco105> rrdns works, but doesn't account for failure
[23:31] <m0zes> is mount.ceph smart enough to retry if a mon doesn't respond?
[23:32] <bilco105> bdonnahue: you mean ceph mon?
[23:32] <bilco105> it needs to be on a unique IP address, different from the other mons
[23:32] <bdonnahue> basically all the ceph services i want going throuhg a vip
[23:33] <bilco105> you can't use a VIP, everything needs unique IP addresses so everything can communicate with each other
[23:33] <bilco105> you should never need to use a VIP
[23:34] <bdonnahue> dude im an idiot.. sorry replace vip with vpn
[23:34] * sherlocked (~watson@ Quit (Quit: Leaving)
[23:34] <bdonnahue> i want the ceph traffic going throguh vpn
[23:34] * bandrus (~brian@207.sub-70-211-69.myvzw.com) Quit (Ping timeout: 480 seconds)
[23:34] <bilco105> m0zes: I don't believe it will retry no, only timeout
[23:34] <bdonnahue> i want to tell ceph to only listen on the tunnel adaptor rather than all interfaces
[23:34] <bdonnahue> previously i got this working with the firewall but that isnt the best solution
[23:34] <bilco105> m0zes: it will cycle through numerous mons if you specify them though
[23:36] <bilco105> bdonnahue: ceph what? osd? mon?
[23:36] <m0zes> bilco105: then maybe I have somewhere I can add a feature. ideally you'd have it do the lookup, if it returns multiple A records, add all of them as mons to the normal mount.
[23:36] <bdonnahue> i want all internal ceph communication to go through a vpn
[23:36] <bilco105> m0zes: maybe. we use puppet for that
[23:36] <bdonnahue> so all the osd intercom and mds and mon intercom to be encrypted
[23:38] <jcsp1> bdonnahue: if you want to use a VPN that's your business ??? ceph just speaks TCP/IP, and you need to make sure the mons IP addresses don't change around.
[23:38] * daniel2_ (~daniel2_@cpe-24-28-6-151.austin.res.rr.com) Quit (Quit: My Mac has gone to sleep. ZZZzzz???)
[23:39] <bilco105> bdonnahue: then specify the cluster_network as the VPN network range, and put that IP address on one of your network adapters.
[23:39] <bdonnahue> jcsp1: i can assure they wont change im just not sure how to configure the daemons what interface to listen on
[23:39] <bdonnahue> or waht ip*
[23:39] <bilco105> bdonnahue: http://ceph.com/docs/master/rados/configuration/network-config-ref/
[23:39] * reed (~reed@75-101-54-131.dsl.static.fusionbroadband.com) has joined #ceph
[23:40] <bilco105> bdonnahue: see public/cluster network under global
[23:40] <bdonnahue> ahh thanks thats what i was looking for
[23:41] <bilco105> Is your ceph cluster network on a shared network?
[23:41] <bdonnahue> yes
[23:41] <bilco105> interesting
[23:41] <jcsp1> we recently had someone get confused about this, so I'll point it out: the cluster network is just for OSD->OSD communications. Your mons still need to be on your public network.
[23:42] <jcsp1> (as do all your clients)
[23:42] <bdonnahue> its in my dev env which is on an open network and used by alot of people .. so i dont really trust that i will have privacy there
[23:42] <bilco105> jcsp1: mons sit in both networks
[23:42] <bdonnahue> jcsp1: thanks for the clarification. I admit that tripped me up
[23:42] <bilco105> or should I say, should be able to see both networks
[23:43] * daniel2_ (~daniel2_@cpe-24-28-6-151.austin.res.rr.com) has joined #ceph
[23:43] * reed (~reed@75-101-54-131.dsl.static.fusionbroadband.com) Quit ()
[23:44] * reed (~reed@75-101-54-131.dsl.static.fusionbroadband.com) has joined #ceph
[23:44] <jcsp1> bilco105: I don't think so. What are you basing that on?
[23:44] <lurbs> The monitors don't actually need to be on either of the public or cluster networks.
[23:44] <bilco105> That OSDs need to be able to connect to mons
[23:44] <bilco105> Yeah, that's why I corrected what I said
[23:44] <bdonnahue> looking at the diagrams i think jcsp1 is correct
[23:45] * oro (~oro@80-219-254-208.dclient.hispeed.ch) has joined #ceph
[23:45] <bdonnahue> nd bilco105 as well
[23:45] <bilco105> sorry, you're right - I'm being dense :)
[23:46] * bandrus (~brian@224.sub-70-211-76.myvzw.com) has joined #ceph
[23:46] <bilco105> forgot for a second that OSDs sit in public as well
[23:51] * reed (~reed@75-101-54-131.dsl.static.fusionbroadband.com) Quit (Quit: Ex-Chat)
[23:51] * reed (~reed@75-101-54-131.dsl.static.fusionbroadband.com) has joined #ceph
[23:52] * reed (~reed@75-101-54-131.dsl.static.fusionbroadband.com) Quit ()
[23:52] * ChrisNBlum (~ChrisNBlu@ Quit (Quit: My Mac has gone to sleep. ZZZzzz???)
[23:53] * bene2 (~ben@nat-pool-bos-t.redhat.com) Quit (Quit: Konversation terminated!)
[23:54] * reed (~reed@75-101-54-131.dsl.static.fusionbroadband.com) has joined #ceph
[23:54] * Deiz (~Jones@5NZAAA0DI.tor-irc.dnsbl.oftc.net) Quit ()
[23:58] * bandrus (~brian@224.sub-70-211-76.myvzw.com) Quit (Ping timeout: 480 seconds)
[23:59] * Izanagi (~Azerothia@tor.de.smashthestack.org) has joined #ceph

These logs were automatically created by CephLogBot on irc.oftc.net using the Java IRC LogBot.