[Linux-HA] Resources are stopped for a few seconds when bringing secondary node out of standby

[Andrew Beekhof]

On Feb 7, 2008, at 2:13 PM, Paul Court wrote:

> When I use crm_standby -U node2 -v false to bring a secondary node  
> out of standby my esources are stopped for a few seconds on the  
> active node and then restarted (still on the active node).

Hi,

The short answer is that the drbd0 most wants to be promoted on  
mysql-01 (despite already being the master on mysql-02) which is  
throwing everything off.
Perhaps the RA should be updating the scores after promotion/demotion  
events.

The long answer is that regardless of what the RA did, the PE is also  
behaving sub-optimally and I've logged a bug for myself so that I'll  
fix it after my vacation

     http://bugs.clusterlabs.org/cgi-bin/bugzilla/show_bug.cgi?id=8

Alas the human brain is exceedingly complex to imitate and the PE  
doesn't always see the "obvious" allocation solution.

>
>
> Also, when running the hb_report program I got the following. (I am  
> running Ubuntu 7.10 but I have relinked /bin/sh to /bin/bash and not  
> "dash".
>
>
> root at mysql-01:~# hb_report -f 12:53 -u root /tmp/hbreport1
> root at mysql-02's password:
> root at mysql-02's password:
> stdin: is not a tty
> root at mysql-02's password:
> /usr/sbin/hb_report: 664: arith: syntax error: "(last+first)/2"
> mysql-02: WARN: couldn't find line for time 1202388780; corrupt log  
> file?
> /usr/sbin/hb_report: 691: arith: syntax error: "maxwait-1"
> press enter to edit the problem description...
>
>
> Paul
> (aka. Gargoyle)
> <hbreport1.tar.gz>_______________________________________________
> Linux-HA mailing list
> Linux-HA at lists.linux-ha.org
> http://lists.linux-ha.org/mailman/listinfo/linux-ha
> See also: http://linux-ha.org/ReportingProblems