Hi Ashley,

The only change I made was increasing the osd_max_backfills from 3 to 10 at first, and when that ended up causing more problems than it helped, it was lowering the setting back down to 3 that took the cluster offline.

I've actually been working on this issue for a week now and my company called in some outside help that walked me through some more intense troubleshooting, and the solution we used was to remove every ceph-mon except one and let it boot up by itself so it wouldn't be looking for a quorum, and that fixed the majority of our problems. It turned out that there was a potential networking issue that sort of popped up and then fixed itself? We weren't sure what exactly happened or how to dig into it once it started working again, but we're sure it had something to do either with the physical NICs or the switches, possibly overloaded with the changes and recovery traffic spiking.

On Tue, Sep 3, 2019 at 9:17 AM Ashley Merrick <singapore@amerrick.co.uk> wrote:
What change did you make in ceph.conf

Id check that hasn't caused an issue first.

---- On Tue, 27 Aug 2019 04:37:15 +0800 nkerns92@gmail.com wrote ----


I have an old ceph 0.94.10 cluster that had 10 storage nodes with one extra management node used for running commands on the cluster. Over time we'd had some hardware failures on some of the storage nodes, so we're down to 6, with ceph-mon running on the management server and 4 of the storage nodes. We attempted deploying a ceph.conf change and restarted ceph-mon and ceph-osd services, but the cluster went down on us. We found all the ceph-mons are stuck in the electing state, I can't get any response from any ceph commands but I found I can contact the daemon directly and get this information (hostnames removed for privacy reasons):

root@<mgmt1>:~# ceph daemon mon.<mgmt1> mon_status
    "name": "<mgmt1>",
    "rank": 0,
    "state": "electing",
    "election_epoch": 4327,
    "quorum": [],
    "outside_quorum": [],
    "extra_probe_peers": [],
    "sync_provider": [],
    "monmap": {
        "epoch": 10,
        "fsid": "69611c75-200f-4861-8709-8a0adc64a1c9",
        "modified": "2019-08-23 08:20:57.620147",
        "created": "0.000000",
        "mons": [
                "rank": 0,
                "name": "<mgmt1>",
                "addr": "[fdc4:8570:e14c:132d::15]:6789\/0"
                "rank": 1,
                "name": "<mon1>",
                "addr": "[fdc4:8570:e14c:132d::16]:6789\/0"
                "rank": 2,
                "name": "<mon2>",
                "addr": "[fdc4:8570:e14c:132d::28]:6789\/0"
                "rank": 3,
                "name": "<mon3>",
                "addr": "[fdc4:8570:e14c:132d::29]:6789\/0"
                "rank": 4,
                "name": "<mon4>",
                "addr": "[fdc4:8570:e14c:132d::151]:6789\/0"

Is there any way to force the cluster back into a quorum even if it's just one mon running to start it up? I've tried exporting the mgmt's monmap and injecting it into the other nodes, but it didn't make any difference.

ceph-users mailing list -- ceph-users@ceph.io
To unsubscribe send an email to ceph-users-leave@ceph.io