[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

RE: [Linux-cluster] Please update cman-kernel and dlm-kernel when themain kernel is updated



Whoa....   

Before you go slapping redhat around I think you should start checking what
updates are
being done and why. If you are running a cluster, it's your responsebility to
make sure
any updates are first "needed" and second, are going to mesh with your
config. 

To answer your question about clustering, the newer clusterring and GFS
packages use the
same underlying subsystems. It's that simple. 

It's done that way to ensure a constant view of "clustered" parts, get
consistent events
etc. Then RHEL 3 way of doing clusters is no more and with good reason. If
they subsystems
don't talk to each other, they tend to step on each other.

That all said, it would be nice if GFS/Cluster updates did come out at or
around the same 
time as kernel updates. I do not use up2date and cannot in my situation so
manual downloads
do it for me. That's ok tho cuz I like to know whats happening on my
clusters.

Finally, this is not the place to complain about such things. If you have a
beef with Redhat
then you should be calling under your support contract and complain there
where an official 
chain of events occurrs...  Your more likely to get your desired result.
Paying customers
get the ear. Anyone (ie non-paying cluster users) can complain here so it's
of much less 
value I think. Your mileage may vary.


Just my 01 cents...


Cheers


Corey

NOTE: My views are MY OWN, NOT REDHAT's  (They may be more kind....)

-----Original Message-----
From: linux-cluster-bounces redhat com
[mailto:linux-cluster-bounces redhat com] On Behalf Of Troels Arvin
Sent: Wednesday, May 24, 2006 7:41 AM
To: linux-cluster redhat com
Subject: [Linux-cluster] Please update cman-kernel and dlm-kernel when
themain kernel is updated

I just ran an up2date on some servers in a RHEL 4 ES cluster, and the kernel
was updated. Now, I tried rebooting one of the nodes in the cluster, and at
least one of the cluster management software daemons didn't properly start,
seemingly due to kernel module mismatches.
Reverting to the older kernel made the cluster node join the cluster again.

This is not impressive; it's actually very disappointing. Red Hat, please
coordinate things.

By the way: Why are special kernel modules even needed by the cluster
software if GFS isn't used?

--
Greetings from Troels Arvin


--
Linux-cluster mailing list
Linux-cluster redhat com
https://www.redhat.com/mailman/listinfo/linux-cluster


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]