https://tracker.ceph.com/https://tracker.ceph.com/favicon.ico2016-04-06T10:56:23ZCeph CephFS - Bug #15399: MDS incarnation get lost after remove filesystem https://tracker.ceph.com/issues/15399?journal_id=687382016-04-06T10:56:23ZJohn Sprayjcspray@gmail.com
<ul></ul><p>Here's a reproducer for the incarnation issue:<br /><a class="external" href="https://github.com/ceph/ceph-qa-suite/tree/wip-15399">https://github.com/ceph/ceph-qa-suite/tree/wip-15399</a></p>
<p>I note that when OSDs construct their objecters they just use the OSDMap epoch as the incarnation. I wonder why the MDS has a per-rank incarnation counter at all? Perhaps we can just remove it and use the MDSMap epoch instead.</p> CephFS - Bug #15399: MDS incarnation get lost after remove filesystem https://tracker.ceph.com/issues/15399?journal_id=687522016-04-06T14:09:57ZGreg Farnumgfarnum@redhat.com
<ul></ul>If:
<ul>
<li>MDSes A and B come up during the same epoch</li>
<li>A becomes active and B becomes standby</li>
<li>A fails</li>
<li>B starts replaying operations</li>
</ul>
<p>Then B needs to have an incarnation which differs from A's. The OSDs are each a distinct daemon entity (unlike the MDSes).</p>
<p>There may be ways to simplify it, though!</p> CephFS - Bug #15399: MDS incarnation get lost after remove filesystem https://tracker.ceph.com/issues/15399?journal_id=687532016-04-06T14:12:17ZGreg Farnumgfarnum@redhat.com
<ul></ul><p>So we could probably reset our network connections with an incarnation based on the last MDSMap where our role changed...I think that should work; maybe it's what you meant.</p> CephFS - Bug #15399: MDS incarnation get lost after remove filesystem https://tracker.ceph.com/issues/15399?journal_id=687642016-04-06T14:51:34ZJohn Sprayjcspray@gmail.com
<ul></ul><p>We only do objecter->set_client_incarnation(incarnation); in MDSRank::init (after we've been assigned an active role)</p>
<p>So epoch should be sufficient (it's always incremented when a rank assignment has changed) as long as we remember to set it when a standby replay MDS is promoted (as well as when MDSRank is initialized).</p> CephFS - Bug #15399: MDS incarnation get lost after remove filesystem https://tracker.ceph.com/issues/15399?journal_id=688222016-04-07T07:02:08ZZheng Yanukernel@gmail.com
<ul></ul><p>I think using the MDSMap epoch as the incarnation is good idea</p> CephFS - Bug #15399: MDS incarnation get lost after remove filesystem https://tracker.ceph.com/issues/15399?journal_id=692222016-04-14T21:36:53ZGreg Farnumgfarnum@redhat.com
<ul><li><strong>Status</strong> changed from <i>New</i> to <i>In Progress</i></li><li><strong>Assignee</strong> set to <i>John Spray</i></li></ul><p>For Jewel: <a class="external" href="https://github.com/ceph/ceph/pull/8484">https://github.com/ceph/ceph/pull/8484</a></p>
<p>But a more comprehensive one (that works with pools shared between FSes) is still in progress.</p> CephFS - Bug #15399: MDS incarnation get lost after remove filesystem https://tracker.ceph.com/issues/15399?journal_id=698362016-04-30T18:22:19ZJohn Sprayjcspray@gmail.com
<ul><li><strong>Status</strong> changed from <i>In Progress</i> to <i>Fix Under Review</i></li></ul><p><a class="external" href="https://github.com/ceph/ceph/pull/8484">https://github.com/ceph/ceph/pull/8484</a></p> CephFS - Bug #15399: MDS incarnation get lost after remove filesystem https://tracker.ceph.com/issues/15399?journal_id=700252016-05-04T10:49:21ZJohn Sprayjcspray@gmail.com
<ul><li><strong>Status</strong> changed from <i>Fix Under Review</i> to <i>Pending Backport</i></li></ul> CephFS - Bug #15399: MDS incarnation get lost after remove filesystem https://tracker.ceph.com/issues/15399?journal_id=700292016-05-04T12:25:04ZNathan Cutlerncutler@suse.cz
<ul><li><strong>Backport</strong> set to <i>jewel</i></li></ul> CephFS - Bug #15399: MDS incarnation get lost after remove filesystem https://tracker.ceph.com/issues/15399?journal_id=700452016-05-04T14:09:54ZNathan Cutlerncutler@suse.cz
<ul><li><strong>Copied to</strong> <i><a class="issue tracker-9 status-3 priority-4 priority-default closed" href="/issues/15732">Backport #15732</a>: jewel: MDS incarnation get lost after remove filesystem </i> added</li></ul> CephFS - Bug #15399: MDS incarnation get lost after remove filesystem https://tracker.ceph.com/issues/15399?journal_id=706152016-05-11T19:23:49ZGreg Farnumgfarnum@redhat.com
<ul><li><strong>Status</strong> changed from <i>Pending Backport</i> to <i>Resolved</i></li></ul>