If revoking of readonly delegations fail, CTDB reschedules a retry immediately. If a node becomes disconnected during this time, CTDB is stuck retrying revoke.
Created attachment 13088 [details] Patch for v4-5, v4-6
Hi Karolin, This is ready for 4.5 and 4.6. Thanks...
(In reply to Martin Schwenke from comment #2) Pushed to autobuild-v4-{6,5}-test.
Pushed to both branches. Closing out bug report. Thanks!
This fix causes a regression starving write requests. Since all the requests are deferred on that node, a read-only request from a different node can cause read-only delegations to be recreated. Thus, wasting the revoke effort and write request has to re-do the revoke of readonly delegations.
Created attachment 13236 [details] Patches for v4-5, v4-6
(In reply to Martin Schwenke from comment #7) Pushed to autobuild-v4-{5,6}-test.
(In reply to Karolin Seeger from comment #8) Pushed to both branches. Closing out bug report. Thanks!