From: Paul H. Hargrove (PHHargrove_at_lbl_dot_gov)
Date: Thu Feb 07 2008 - 17:42:27 PST
Jerome,
I am fairly certain you are seeing the ill-effects of "prelinking".
We have an FAQ on this issue:
http://mantis.lbl.gov/blcr/doc/html/FAQ.html#prelink
Let us know if that proves not to be the source of your problem and
we'll see how we can help
-Paul
Jerome wrote:
> Hi all
>
> i'm just beginning to use BLCR library for my own cluster, in the case
> of package that's dont include chekpointing avaibility.
> As to understand how BLCR run, i'm just doing a dummy program as
> "hello world" with a sleep command to have time to do a checkpoint .-)
>
> I run this program on mi cluster's master node and on the nodes.
> But i'v notice that when i do a checkpoint on the master node, i'v got
> an horrible "Segmentation Fault" restarting it on a node. And the
> master and nodes have the same kernel version, the same libraries.
> What i have to do te detect from where comes the problem?
>
> Best regards.
>
>
--
Paul H. Hargrove PHHargrove_at_lbl_dot_gov
Future Technologies Group
HPC Research Department Tel: +1-510-495-2352
Lawrence Berkeley National Laboratory Fax: +1-510-486-6900