Re: Partitioned cluster question (reboot during lost quorum)
- From: Hoff Hoffman <hoff-remove-this@xxxxxx>
- Date: Wed, 19 Apr 2006 23:15:07 GMT
JF Mezei wrote:
So the gist of it is to set every node'S EXPECTED vote to ensure that
they cannot boot alone if disconnected from the ethernet. And manually
set it to "1" TEMPORARILY if you really need a node to boot alone/first
before allowing others to join.
Y'all are trying to get "creative" again, eh?
If you want to boot outside the cluster, then I tend to prefer to avoid enabling NISCS_LOAD_PEA0 and I don't load VAXCLUSTER.
(consider the case when a node is known
to have the valid copy of shadow disk members, so you want that one to
boot first and mount them first so they become source of future copy
operations isntead of targets).
Consider the last time this has happened, too. I've been at this for over a decade, and I can't recall ever having run into this case. As a rule, the node with the valid copy will have the highest instantiation, and will be the source for the shadow copy operation, or there have been other "creative" activities with the shadow set members. When you get (incautiously) "creative" with shadowing -- just like with the cluster vote-related system parameters -- you can end up corrupting a disk.
Assuming I could not simply take the other shadow set member disks off-line (which is what I try do when I get "creative" with shadow sets, in addition to managing the MSCPMOUNT.COM or startup procedures around the MOUNT operations), I'd confirm the bootstrap state, and I'd confirm that all other nodes are down (hard), and I'd then drop the EXPECTED_VOTES setting to match the local voting node's VOTES setting during conversational bootstrap, and I'd then (try to) remember to then clear WRITESYSPARAMS, and then and only then complete the bootstrap.
.
- Follow-Ups:
- Re: Partitioned cluster question (reboot during lost quorum)
- From: JF Mezei
- Re: Partitioned cluster question (reboot during lost quorum)
- References:
- Partitioned cluster question (reboot during lost quorum)
- From: JF Mezei
- Re: Partitioned cluster question (reboot during lost quorum)
- From:
- Re: Partitioned cluster question (reboot during lost quorum)
- From: JF Mezei
- Re: Partitioned cluster question (reboot during lost quorum)
- From: Hoff Hoffman
- Re: Partitioned cluster question (reboot during lost quorum)
- From: JF Mezei
- Partitioned cluster question (reboot during lost quorum)
- Prev by Date: Re: Strange behavior of DEFINE with equal-sign
- Next by Date: Re: OT: Sparc not dead yet
- Previous by thread: Re: Partitioned cluster question (reboot during lost quorum)
- Next by thread: Re: Partitioned cluster question (reboot during lost quorum)
- Index(es):
Relevant Pages
|