aaicraid battery != Ok = Server stone dead (allmost) forever

Robert Lindgren robert at lunarpages.com
Tue Dec 10 10:03:00 CST 2002


Well for some reason I really don't believe that, all of our servers
have the tg3 nics and doesn't show the same locking problem. There isn't
anything specific with the server locking compared to the others, I mean
like the netload is the same for all.

How is it possible that the server is slow right after a reboot if it
would be the tg3 problem, and starts to work fine when the battery state
changes to Ok?

/Rob

On Mon, 2002-12-09 at 23:29, Steve_Boley at dell.com wrote:
> Look elsewhere, that's not what is locking your system.  Probably related to
> tg3 module for broadcom nics and the locking it has been causing.
> Steve
> 
> -----Original Message-----
> From: Robert Lindgren [mailto:robert at lunarpages.com]
> Sent: Monday, December 09, 2002 7:00 AM
> To: Linux-PowerEdge at exchange.dell.com
> Subject: aaicraid battery != Ok = Server stone dead (allmost) forever
> 
> 
> Hi all,
> 
> I have some problems with an 2650, at any arbitrary point in time the
> server just stops to respond. I believe this has something to do with
> the battery on the raid controller, but since I'm not able to login to
> the machine when this happens I'm not able to state that it is related
> to the battery though.
> 
> Sometimes I'm able to get the password prompt back from the server but
> I'm not able to login to it, seems like it does allot of i/o. When the
> server is rebooted the problem still persist until I get this message in
> the system log:
> 
> kernel: aacraid:Battery charge is now OK
> 
> So my question(s) is can this problem be related to a faulty battery on
> the raid controller? If the first question is true, how can this bring
> the server to a state where it's not possible to do ANY i/o related
> stuff, like logging in to the system. Sure the cache on the controller
> is disabled but, never seen something like this on a (sorry curse words
> will follow) Compaq or Sun server when the battery has been discharged.
> 
> Best regards
-- 
Robert Lindgren <robert at lunarpages.com>




More information about the Linux-PowerEdge mailing list