Re: Monitors stuck in "electing"

[Travis Rhoden <trhoden@xxxxxxxxx>]

I bumped debug mon and debug ms up on one of the monitors (ceph0), and this is what I see:

2014-03-25 16:02:19.273406 7fa5c53fc700  5 mon.ceph0@0(electing).elector(35) election timer expired
2014-03-25 16:02:19.273447 7fa5c53fc700  5 mon.ceph0@0(electing).elector(35) start -- can i be leader?
2014-03-25 16:02:19.273528 7fa5c53fc700  1 mon.ceph0@0(electing).elector(35) init, last seen epoch 35
2014-03-25 16:02:19.273543 7fa5c53fc700  1 -- 10.10.30.0:6789/0 --> mon.1 10.10.30.1:6789/0 -- election(b3f38955-4321-4850-9ddb-3b09940dc951 propose 35) v4 -- ?+0 0x251bd80
2014-03-25 16:02:19.273569 7fa5c53fc700  1 -- 10.10.30.0:6789/0 --> mon.2 10.10.30.2:6789/0 -- election(b3f38955-4321-4850-9ddb-3b09940dc951 propose 35) v4 -- ?+0 0x251bb40
2014-03-25 16:02:20.284459 7fa5c4bfb700  1 -- 10.10.30.0:6789/0 <== mon.2 10.10.30.2:6789/0 312 ==== election(b3f38955-4321-4850-9ddb-3b09940dc951 propose 35) v4 ==== 537+0+0 (4036841703 0 0) 0x2515480 con 0x1df0c60
2014-03-25 16:02:20.284524 7fa5c4bfb700  5 mon.ceph0@0(electing).elector(35) handle_propose from mon.2
2014-03-25 16:02:24.273726 7fa5c53fc700  5 mon.ceph0@0(electing).elector(35) election timer expired

That just repeats...

On Tue, Mar 25, 2014 at 11:48 AM, Travis Rhoden <trhoden@xxxxxxxxx> wrote:

Just to emphasize that I don't think it's clock skew, here is the NTP state of all three monitors:

# ansible ceph_mons -m command -a "ntpq -p" -kK
SSH password:
sudo password [defaults to SSH password]:
ceph0 | success | rc=0 >>
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
*controller-10g  198.60.73.8      2 u   43   64  377    0.236    0.057   0.097

ceph1 | success | rc=0 >>
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
*controller-10g  198.60.73.8      2 u   39   64  377    0.273    0.035   0.064

ceph2 | success | rc=0 >>
     remote           refid      st t when poll reach   delay   offset  jitter
==============================================================================
*controller-10g  198.60.73.8      2 u   30   64  377    0.201   -0.063   0.063

I think they are pretty well in synch.

 - Travis

On Tue, Mar 25, 2014 at 11:09 AM, Travis Rhoden <trhoden@xxxxxxxxx> wrote:

Hello,

I just deployed a new Emperor cluster using ceph-deploy 1.4.  All went very smooth, until I rebooted all the nodes.  After reboot, the monitors no longer form a quorum.

I followed the troubleshooting steps here: http://ceph.com/docs/master/rados/troubleshooting/troubleshooting-mon/

Specifically, I"m in the stat described in this section: http://ceph.com/docs/master/rados/troubleshooting/troubleshooting-mon/#most-common-monitor-issues

The state for all the monitors is "electing".  The docs say this is most likely clock skew, but I do have all nodes synch'd with NTP.  I've confirmed this multiple times.  I've also confirmed the monitors can reach each other (by telneting to IP:PORT, and I can see established connections via netstat).

I'm baffled.

here is a sample mon_status output:

root@ceph0:~# ceph daemon mon.ceph0 quorum_status
{ "election_epoch": 31,
  "quorum": [],
  "quorum_names": [],
  "quorum_leader_name": "",
  "monmap": { "epoch": 2,
      "fsid": "XXX", (redacted)
      "modified": "2014-03-24 14:35:22.332646",
      "created": "0.000000",
      "mons": [
            { "rank": 0,
              "name": "ceph0",
              "addr": "10.10.30.0:6789\/0"},
            { "rank": 1,
              "name": "ceph1",
              "addr": "10.10.30.1:6789\/0"},
            { "rank": 2,
              "name": "ceph2",
              "addr": "10.10.30.2:6789\/0"}]}}

They all look identical to that.

Any ideas what I can look at besides NTP?  The docs really stress that it should be clock skew, so I'll keep looking at that...

 - Travis

_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com