most of the reason cause Mon down is big log
i have set debug paxos = 0, and i am watching now.
before set this,
# ceph daemon mon.ceph01-vm config get debug_mon
{ "debug_mon": "1\/5"}
# ceph daemon mon.ceph01-vm config get debug_ms
{ "debug_ms": "0\/5"}
# ceph daemon mon.ceph01-vm config get debug_paxos
{ "debug_paxos": "1\/5"}
after set this
# ceph daemon mon.ceph01-vm config get debug_mon
{ "debug_mon": "1\/5"}
# ceph daemon mon.ceph01-vm config get debug_ms
{ "debug_ms": "0\/5"}
# ceph daemon mon.ceph01-vm config get debug_paxos
{ "debug_paxos": "0\/0"}
2014-08-24 18:58 GMT+07:00 Joao Eduardo Luis <joao.luis at inktank.com>:
> On 08/24/2014 01:57 AM, debian Only wrote:
>
>> this is happen i use *ceph-deploy create ceph01-vm ceph02-vm ceph04-vm
>> *to create 3 Mons member.
>>
>> now every 10 hours, one Mon will down. every time have this error,
>> some time the hardisk have enough space left,such as 30G.
>>
>> i deployed Ceph before, only create one Mon at first step *ceph-deploy
>> create ceph01-vm , and then ceph-deploy mon add ceph02-vm, *not meet
>>
>> this problem.
>>
>> i do not know why ?
>>
>
> Your monitor shutdown because the disk the monitor is sitting on has
> dropped to (or below) 5% of available disk space. This is meant to prevent
> the monitor from running out of disk space and be unable to store critical
> cluster information. 5% is a rough estimate, which may be adequate for
> some disks, but may be either too small or too large for small disks and
> large disks respectively. This value can be adjusted if you feel like you
> need to, using the 'mon_data_avail_crit' option (which defaults to 5, as in
> 5%, but can be adjusted to whatever suits you best).
>
> The big problem here however seems to be that you're running out of space
> due to huge monitor logs. Is that it?
>
> If so, I would ask you to run the following commands and share the results:
>
> ceph daemon mon.* config get debug_mon
> ceph daemon mon.* config get debug_ms
> ceph daemon mon.* config get debug_paxos
>
> -Joao
>
>
>> 2014-08-23 10:19:43.910650 7f3c0028c700 0
>> mon.ceph01-vm at 1(peon).data_health(56) *update_stats avail 5% total
>> 15798272 used 12941508 avail 926268*
>>
>> 2014-08-23 10:19:43.910806 7f3c0028c700 -1
>> mon.ceph01-vm at 1(peon).data_health(56) reached critical levels of
>> available space on local monitor storage -- shutdown!
>> 2014-08-23 10:19:43.910811 7f3c0028c700 0 ** Shutdown via Data Health
>> Service **
>> 2014-08-23 10:19:43.931427 7f3bffa8b700 1
>> mon.ceph01-vm at 1(peon).paxos(paxos active c 15814..16493) is_readable
>> now=2014-08-23 10:19:43.931433 lease_expire=2014-08-23 10:19:45.989585
>> has v0 lc 16493
>> 2014-08-23 10:19:43.931486 7f3bfe887700 -1 mon.ceph01-vm at 1(peon) e2 ***
>> Got Signal Interrupt ***
>> 2014-08-23 10:19:43.931515 7f3bfe887700 1 mon.ceph01-vm at 1(peon) e2
>> shutdown
>> 2014-08-23 10:19:43.931725 7f3bfe887700 0 quorum service shutdown
>> 2014-08-23 10:19:43.931730 7f3bfe887700 0
>> mon.ceph01-vm at 1(shutdown).health(56) HealthMonitor::service_shutdown 1
>> services
>> 2014-08-23 10:19:43.931735 7f3bfe887700 0 quorum service shutdown
>>
>>
>>
>> 2014-08-22 21:31 GMT+07:00 debian Only <onlydebian at gmail.com
>> <mailto:onlydebian at gmail.com>>:
>>
>>
>> this time ceph01-vm down, no big log happen , other 2 ok. do not
>> what's the reason, this is not my first time install Ceph. but
>> this is first time i meet that mon down again and again.
>>
>> ceph.conf on each OSDs and MONs
>> [global]
>> fsid = 075f1aae-48de-412e-b024-b0f014dbc8cf
>> mon_initial_members = ceph01-vm, ceph02-vm, ceph04-vm
>> mon_host = 192.168.123.251,192.168.123.252,192.168.123.250
>> auth_cluster_required = cephx
>> auth_service_required = cephx
>> auth_client_required = cephx
>> filestore_xattr_use_omap = true
>>
>> rgw print continue = false
>> rgw dns name = ceph-radosgw
>> osd pool default pg num = 128
>> osd pool default pgp num = 128
>>
>>
>> [client.radosgw.gateway]
>> host = ceph-radosgw
>> keyring = /etc/ceph/ceph.client.radosgw.keyring
>> rgw socket path = /var/run/ceph/ceph.radosgw.gateway.fastcgi.sock
>> log file = /var/log/ceph/client.radosgw.gateway.log
>>
>>
>> 2014-08-22 18:15 GMT+07:00 Joao Eduardo Luis <joao.luis at inktank.com
>> <mailto:joao.luis at inktank.com>>:
>>
>>
>> On 08/22/2014 10:21 AM, debian Only wrote:
>>
>> i have 3 mons in Ceph 0.80.5 on Wheezy. have one RadosGW
>>
>> when happen this first time, i increase the mon log device.
>> this time mon.ceph02-vm down, only this mon down, other 2
>> is ok.
>>
>> pls some one give me some guide.
>>
>> 27M Aug 22 02:11 ceph-mon.ceph04-vm.log
>> 43G Aug 22 02:11 ceph-mon.ceph02-vm.log
>> 2G Aug 22 02:11 ceph-mon.ceph01-vm.log
>>
>>
>> Depending on the debug level you set, and depending on which
>> subsystems you set a higher debug level, the monitor can spit
>> out A LOT of information in a short period of time. 43GB is
>> nothing compared to some 100+ GB logs I've had churn through in
>> the past.
>>
>> However, I'm not grasping what kind of help you need. According
>> to your 'ceph -s' below the monitors seem okay -- all are in,
>> health is OK.
>>
>> If you issue is with having that one monitor spitting out
>> humongous amounts of debug info here's what you need to do:
>>
>> - If you added one or more 'debug <something> = X' to that
>> monitor's ceph.conf, you will want to remove them so that in a
>> future restart the monitor doesn't start with non-default debug
>> levels.
>>
>> - You will want to inject default debug levels into that one
>> monitor.
>>
>> Depending on what debug levels you have increased, you will want
>> to run a version of "ceph tell mon.ceph02-vm injectargs
>> '--debug-mon 1/5 --debug-ms 0/5 --debug-paxos 1/5'"
>>
>> -Joao
>>
>>
>> # ceph -s
>> cluster 075f1aae-48de-412e-b024-__b0f014dbc8cf
>>
>> health HEALTH_OK
>> monmap e2: 3 mons at
>> {ceph01-vm=192.168.123.251:__6789/0,ceph02-vm=192.168.123._
>> _252:6789/0,ceph04-vm=192.168.__123.250:6789/0
>>
>> <http://192.168.123.251:6789/0,ceph02-vm=192.168.123.252:
>> 6789/0,ceph04-vm=192.168.123.250:6789/0>
>> <http://192.168.123.251:6789/__0,ceph02-vm=192.168.123.252:_
>> _6789/0,ceph04-vm=192.168.123.__250:6789/0
>>
>> <http://192.168.123.251:6789/0,ceph02-vm=192.168.123.252:
>> 6789/0,ceph04-vm=192.168.123.250:6789/0>>},
>>
>> election epoch 44, quorum 0,1,2 ceph04-vm,ceph01-vm,ceph02-vm
>> mdsmap e10: 1/1/1 up {0=ceph06-vm=up:active}
>> osdmap e145: 10 osds: 10 up, 10 in
>> pgmap v4394: 2392 pgs, 21 pools, 4503 MB data, 1250
>> objects
>> 13657 MB used, 4908 GB / 4930 GB avail
>> 2392 active+clean
>>
>>
>> /2014-08-22 02:06:34.738828 7ff2b9557700 1
>>
>> mon.ceph02-vm at 2(peon).paxos(__paxos active c 9037..9756)
>>
>> is_readable
>> now=2014-08-22 02:06:34.738830 lease_expire=2014-08-22
>> 02:06:39.701305
>> has v0 lc 9756/
>> /2014-08-22 02:06:36.618805 7ff2b9557700 1
>>
>> mon.ceph02-vm at 2(peon).paxos(__paxos active c 9037..9756)
>>
>> is_readable
>> now=2014-08-22 02:06:36.618807 lease_expire=2014-08-22
>> 02:06:39.701305
>> has v0 lc 9756/
>> /2014-08-22 02:06:36.620019 7ff2b9557700 1
>>
>> mon.ceph02-vm at 2(peon).paxos(__paxos active c 9037..9756)
>>
>> is_readable
>> now=2014-08-22 02:06:36.620021 lease_expire=2014-08-22
>> 02:06:39.701305
>> has v0 lc 9756/
>> /2014-08-22 02:06:36.620975 7ff2b9557700 1
>>
>> mon.ceph02-vm at 2(peon).paxos(__paxos active c 9037..9756)
>>
>> is_readable
>> now=2014-08-22 02:06:36.620977 lease_expire=2014-08-22
>> 02:06:39.701305
>> has v0 lc 9756/
>> /2014-08-22 02:06:36.629362 7ff2b9557700 0
>> mon.ceph02-vm at 2(peon) e2
>>
>> handle_command mon_command({"prefix": "mon_status",
>> "format": "json"} v
>> 0) v1/
>> /2014-08-22 02:06:36.633007 7ff2b9557700 0
>> mon.ceph02-vm at 2(peon) e2
>> handle_command mon_command({"prefix": "status", "format":
>> "json"} v 0) v1/
>> /2014-08-22 02:06:36.637002 7ff2b9557700 0
>> mon.ceph02-vm at 2(peon) e2
>>
>> handle_command mon_command({"prefix": "health", "detail":
>> "", "format":
>> "json"} v 0) v1/
>> /2014-08-22 02:06:36.640971 7ff2b9557700 0
>> mon.ceph02-vm at 2(peon) e2
>>
>> handle_command mon_command({"dumpcontents": ["pgs_brief"],
>> "prefix": "pg
>> dump", "format": "json"} v 0) v1/
>> /2014-08-22 02:06:36.641014 7ff2b9557700 1
>>
>> mon.ceph02-vm at 2(peon).paxos(__paxos active c 9037..9756)
>>
>> is_readable
>> now=2014-08-22 02:06:36.641016 lease_expire=2014-08-22
>> 02:06:39.701305
>> has v0 lc 9756/
>> /2014-08-22 02:06:37.520387 7ff2b9557700 1
>>
>> mon.ceph02-vm at 2(peon).paxos(__paxos active c 9037..9757)
>>
>> is_readable
>> now=2014-08-22 02:06:37.520388 lease_expire=2014-08-22
>> 02:06:42.501572
>> has v0 lc 9757/
>>
>>
>>
>> _________________________________________________
>> ceph-users mailing list
>> ceph-users at lists.ceph.com <mailto:ceph-users at lists.ceph.com>
>> http://lists.ceph.com/__listinfo.cgi/ceph-users-ceph.__com
>>
>> <http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com>
>>
>>
>>
>> --
>> Joao Eduardo Luis
>> Software Engineer | http://inktank.com | http://ceph.com
>>
>>
>>
>>
>
> --
> Joao Eduardo Luis
> Software Engineer | http://inktank.com | http://ceph.com
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ceph.com/pipermail/ceph-users-ceph.com/attachments/20140824/c9a2e0c4/attachment.htm>
