[SSI-devel] Re: Fw: drbd

[Kilian C. <kil...@st...>]

Jaideep Dharap wrote:
> I would suggest to check 3 things:
> 1. Make sure you have "chard" and the nodes listed "1:2" (for nodes 1
> and 2) in the fstab.
>     e.g. /dev/nbd/1 <mntpt> ext3    default,chard,node=1:2  0       0

I have:
/dev/nbd/0    /    ext3    chard,defaults,node=1:5    1       1

> 2. Make sure /etc/init.d/SSIfailover is using the right device for failover
> Notice that the findfs line is commented out.

That's ok too:
node1# grep DEVICE= /etc/init.d/SSIfailover 
        #DEVICE=`/sbin/findfs $FSDEV`
        DEVICE=/dev/nbd/0

> 3. Also could you check the console output during the bootup on second
> node and see if you see the message
>   "Enabled root failover". This should indicate whether the root
> failover has been enabled in the first place.

Ah, I've not this one. But, it seems failover is enabled on secondary node:

node1# onnode 5 /etc/init.d/SSIfailover status
Root failover enabled - Takeover node

But I can't see it in dmesg.
I have "SSIfailover     initnode        Y" in /etc/rc.nodeinfo.
What can I check else to ensure failover is enabled at boot?

> If you confirm all the above, I fear the issue might be a Debian-OpenSSI
> issue. As Aneesh mentioned, even normal
> failover has notbeen tested with Debian. So we might have to look into
> that :-).

Whatever, thanks a lot for your support.

-- 
Kilian CAVALOTTI | GPGKeyId: 0xD657340C
BOFH excuse #309:
firewall needs cooling