|From:||Andrew Atrens <atrens@xxxxxxxxxxxxxxxxxx>|
|Date:||Wed, 1 Sep 2004 22:13:01 -0400|
On September 1, 2004 07:41 pm, YONETANI Tomokazu wrote: > Hi. > > On Wed, Sep 01, 2004 at 02:40:51PM -0400, Andrew Atrens wrote: > > Andrew Atrens wrote: > > > George Georgalis wrote: > > >> On Wed, Sep 01, 2004 at 11:55:18AM -0400, Andrew Atrens wrote: > > >>>compiler error. When running the same build without 'nice' > > >>>it invariably works without error. This behaviour is kind of > > >>>random in that it's always a different file that fails, but > > >>>predictable in that buildworld 'always' fails at some point. > > >> > > >> have you run memtest86 lately? cpu too hot? > > >> > > >> // George > > > > > > Reverting to Aug 27th kernel fixes it ... > > > > Whew... back to my day job ;) ;) > > Do you have a non-ufs filesystem(e.g. nullfs, nfs, ...) mounted > and extensively accessed on your computer? This box is both an nfs client (about 20 static mount points) and a server (about 5 clients, several mount points), and I'm running amd so it's automounting on /net a lot, especially during many clearcase operations. Not sure if 'cleartool lsact -l' does anything with /net but it is possible. I'm thinking that locking in the linuxulator is at the root of this ... Andrew.