At Thu, 21 May 2009 14:30:42 -0400 CentOS mailing list <centos at centos.org> wrote: > > > On Thu, 2009-05-21 at 20:07 +0200, Ralph Angenendt wrote: > > Lanny Marcus wrote: > > > It would have been helpful, if the error message told you which system > > > software. :-) The Upgrade from 5.2 to 5.2 seems to have been > > > problematic for some people on this list. When you did the Upgrade, > > > did you follow this sequence, per the CentOS 5.3 Release Notes? > > > yum clean all && yum update glibc\* && yum update > > > If not, that may or may not have anything to do with the reboots and > > > error messages you have received. > > > > No, it would not have to do anything with the spontaneous reboots. The > > problem with glibc only concerns rpm, as the release notes clearly state. > > Since nobody else mentioned it to th OP, ... > > Let us not forget that often the hardware chooses to act up around the > same time that some kind of (software) upgrade is performed. I've wasted > a lot of time in the past *assuming* that because the hardware was > rock-solid in the past, it must have been some change (I made) to the > software. > > I suggest running diagnostics, or manually re-seating everything > (especially if you had occasion to move the unit or open it recently). > > Memory used to age, does it still? Memtest*86 might be in order. Not so much as age, but contacts corrode and lots of components change values as they are heated and cooled (expand or shrink). And yes, heating and cooling causes small dimensional changes in connectors -- this can (over time) work connectors loose. It is possible that software is *partly* to blame, if only because the software changes might cause little used bits (litterally bits!) of hardware (eg memory) to be used more than they were. You'd never know about bad memory if it is not actually used. I used to have a K6500 that was 'fussy' about a specific stick of 128meg of RAM. The machine was perfectly fine, until I did backups, which involved writing a very large tar file to a removable disk. The disk I/O would fail. There was nothing wrong with the disk(s) or the software. It was the extra 128meg DIMM. And there was nothing really wrong with the DIMM either. It was just a wee bit too slow for the K6500. (eg it was PC99.99999 instead of PC100 or something like that). Only certian sorts of memory accesses would fail. > > > > > Ralph > > <snip sig stuff> > > HTH -- Robert Heller -- 978-544-6933 Deepwoods Software -- Download the Model Railroad System http://www.deepsoft.com/ -- Binaries for Linux and MS-Windows heller at deepsoft.com -- http://www.deepsoft.com/ModelRailroadSystem/