[Linux-cluster] Suggestion for backbone network maintenance

Gianluca Cecchi gianluca.cecchi at gmail.com
Wed Oct 7 15:03:50 UTC 2009


Hello,
cluster rh el 5.3 with 2 nodes and a quorum disk with heuristics. The nodes
are in different sites.
At this moment inside cluster.conf I have this:

        <quorumd device="/dev/mapper/mpath6" interval="5" label="oraquorum"
log_facility="local4" log_level="7" tko="16" votes="1">
                <heuristic interval="2" program="ping -c1 -w1 10.4.5.250"
score="1" tko="20"/>
        </quorumd>

there is a planning for backbone network maintenance and I'm gong to have
interruption on backbone switches in both sides, so that the gw will be not
reachable in different time windows.....
I would like to "downgrade" the cluster to a sort of standalone server, to
prevent scenario of shutdown abort of oracle, due to ping-pong of
service....
and without stopping service itself.
Only thing is that when the network link will be down, the service will be
down too (accettable by users), but without impact to the backend db on the
cluster.
My idea is to temporarily extend tko inside heuristic, so that the
count-down for loss of quorum will not arrive at a end during downtime.
So that if overall planned maintenance is 2h I will put  something like
tko=1500 on the fly.

So my planned commands would be, where node1 is providing service and node2
is standby:

node2
- shutdown -h now (to simplify things...)

node1
- change to cluster.conf incrementing version number and putting tko=1500
- ccs_tool update /etc/cluster/cluster.conf
- cman_tool version -r <new_version>

the new tko will be dynamically in place?
If so, wait for maintenance completion and then

node1
- change to cluster.conf incrementing version number and putting tko=20
again
- ccs_tool update /etc/cluster/cluster.conf
- cman_tool version -r <new_new_version>

node2
power on
it will keep the new configuration without downsides (correct?)

Any hints or comments?

Thanks in advance,
Gianluca
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/linux-cluster/attachments/20091007/4d94cf50/attachment.htm>


More information about the Linux-cluster mailing list