Whats weird, is i have not have seen this before, this is the first
time this has happen to me where the primary does not go down.... so if
i use stonith, and the primary comes bcak up does it switch back then?
Thanks for replying
Rob Morin
Dido Internet Inc.
Montreal,Canada
http://www.dido.ca
514-990-4444
Dejan Muhamedagic wrote:
Hi,
On Wed, May 07, 2008 at 10:08:42AM -0400, Rob Morin wrote:
Hello all....
Last night i was performing a test of the fail over, to do the test i
simply brought down eth0(used for replication and heartbeat, eth1 is
public stuff) now the secondary came up pretty quick, however the primary
also stayed up??
It's called split-brain. Why should the first node (primary) come
down? If you want that to happen deploy stonith. There's
documentation at linux-ha.org about split-brain and why it's bad
and how to prevent it from disrupting your services or toasting
your data.
Thanks,
Dejan
I had to do a /etc/init.d/heartbeat stop in order for the
primary to come down, then the secondary machine was ok except for a
couple mysql issues..... when i restarted heartbeat on the primary
services went back to the primary, but drbd was not connect and i saw a
split brain in the logs.....
Any ideas as to why the primary did not come down? what should i be
checking in the logs...
I am using an older heartbeat & drbd as we stick with whatever Debian
uses....
heartbeat 2.0.7
drbd 8.0.7
Thanks to all
--
Rob Morin
Dido Internet Inc.
Montreal,Canada
http://www.dido.ca
514-990-4444
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
_______________________________________________
Linux-HA mailing list
[email protected]
http://lists.linux-ha.org/mailman/listinfo/linux-ha
See also: http://linux-ha.org/ReportingProblems
