Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Il 01 marzo 2017 23:13:38 CET, "Sean M. Pappalardo" <spappalardo at renegadetech.com> ha scritto: > > >On 03/01/2017 12:41 AM, Roland Kammerer wrote: >>> On 02/28/2017 12:44 AM, Roland Kammerer wrote: >>>> Yes it uses dual-primary, only during the migration. From what we >saw it >>>> works reasonably well in these scenarios. ... >I was just looking for more on what you meant by "reasonably well." > >Did that mean "live migration works sometimes, but when it fails, the >VM I am running a small pve 4 cluster (3 nodes) with drbd9 since 1 year or so. Never had a single live migration failed. My vm are not wite intensive, anyway. >just stays where it is" or more like "sometimes live migration fails >and >you get split-brain or lose the contents of the DRBD resource when it >does"? Never had this problems, only occasional out-of-sync upon reboot on the third node, which is there only for quorum and data redundance (no vms on it, and no battery backup controller). Resolved manually, disconnecting and reconnecting the resource in various ways. I would not recommend anyway to run it in production, due to current instability of drbdmange tool behaviour . rob