ikedaj at intellilink
Nov 1, 2007, 6:52 PM
Post #2 of 2
(195 views)
Permalink
|
> hi, all
>
> I have a cluster about 3-4 node. When I try to use command "hb_delnode",
it
> seems nothing happen.
> This is my configuration file ha.cf:
> keepalive 125ms
> deadtime 500ms
> warntime 250ms
> bcast enet0 enet1
> auto_failback off
> crm on
> autojoin any
>
>
> By the way, I find some information in hb_delnode.in:
> # This script will only work for a two machine setup...
> # More than that and you need to vote, or something...
>
> Does it mean that I can't use "hb_delnode" to del node online? What can I
> do?
I think it's needed to stop the node which you want to delete before calling
"hb_delnode".
If you call "hb_delnode" before the node stops, There might be a message
like this in ha-log.
WARN: deletion failed: node <node name> is not dead
"online" means that you can delete the node while other nodes are alive,
doesn't it?
Here is my operation as an example;
1) stop the node, for instance "node-a".
node-a is a target node which I want to delete here.
# service heartbeat stop
2) call "hb_delnode" on all surviving nodes.
# hb_delnode node-a
You can find the following message in ha-log if "hb_delete" works well.
info: Removing node node-a from configuration.
I usually check "/var/lib/heartbeat/hostcache", too.
node-a should be deleted at this time from this file.
3) call "cibadmin" on DC.
There remains "node-a" on crm_mon as OFFLINE until calling cibadmin.
# cibadmin -D -o nodes -X '<node id="...uuid..." uname="node-a">'
# cibadmin -D -o status -X '<node_state id="...uuid..." uname="node-a">'
That's all for now.
If there is something wrong, please point it out.
There is a small bug in 2.1.2,
But Andrew has already fixed it.
You can take this.
http://hg.linux-ha.org/dev/rev/f1a775fd0069
Thanks,
Junko
_______________________________________________
Linux-HA mailing list
Linux-HA [at] lists
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
|
