Hi, Andrew,
The workload for LAM/MPI is about the same. The only thing changed
lately is the kernel, that's why we are wondering if we should recompile
LAM or our applications.
When upgrading kernel, does the system header files (those in
/usr/include) change ?
Probably not, from 2.6.9-34 to 2.6.9-42.
We've been having this lamnodes hanging problem for a while, but roughly
in a rate of once a month. Lately it increases to once in less than a
week.
Regards,
Lily
-----Original Message-----
From: Andrew Friedley [mailto:afriedle_at_[hidden]]
Sent: Thursday, October 26, 2006 9:49 AM
To: General LAM/MPI mailing list
Subject: Re: LAM: Do we need to recompile LAM and applications after
weupgrade the linux kernel ?
Lily Li wrote:
>
> Hello, everyone,
>
> It seems that after we upgrade the linux kernel from 2.6.9-34.Elsmp to
> 2.6.9-42.Elsmp for our Pentium III cluster, we start having a higher
> rate of lamd hanging problem on the headnode. The lamd will not
> response to the command "lamnodes" after the LAM is booted and used
> for couple of days.
>
> The question is : do we need to recompile/link the LAM and the
> applications after we upgrade the linux kernel ?
I'm not sure why lamd would hang, but I certainly wouldn't expect
changing the kernel to cause it.. especially when both are 2.6.9.
Exactly how often was lamd hanging before, and how often is it hanging
now?
I'm wondering if something else on the system is causing it.. Does
changing back to the old kernel take you back to the old lamd behavior?
Has your LAM/MPI workload changed? Anything else?
Andrew
|
