08.01.2020 17:30, Achim Leitner пишет:
I can't see what is causing this behaviour.
Am 2020-01-11 17:38, schrieb Andrei Borzenkov:
primitive VM-Step8a Delay \
params startdelay=10 stopdelay=2 mondelay=0 \
op start interval=0 timeout=40s \
op stop interval=0 timeout=30s
08.01.2020 17:30, Achim Leitner пишет:
> Hi,
>
> some progress on this issue:
>
> Am 20.12.19 um 13:37 schrieb Achim Leitner:
>> After pacemaker restart, we have Transition 0 with the DRBD actions,
>> followed 4s later with Transition 1 including all VM actions with
>> correct ordering. 32s
Hi,
some progress on this issue:
Am 20.12.19 um 13:37 schrieb Achim Leitner:
> After pacemaker restart, we have Transition 0 with the DRBD actions,
> followed 4s later with Transition 1 including all VM actions with
> correct ordering. 32s later, there is Transition 2 with fewer actions,
> but
Hi all,
TL;DR: Apparently, both the order of adding resources to the CIB and
the number of CPU cores on a pacemaker node might impact advisory
(optional) startorder constraints. This makes it difficult to configure
them for full cluster restarts, e.g. after a power outage, when the
number of