From: Stefan H. <sh...@lu...> - 2006-07-03 20:26:28
|
Am Montag, den 03.07.2006, 10:33 +0200 schrieb Daniel Tombeil: Hi Daniel, > > can anyone help me, please? My systems run very dirty and I cannot use > > any 2.6.17.x kernels. Am I alone with this problem? Under 2.6.16.20 I > > never saw such a kernel panic. > > I had similar looking panics and in my case it was a problem with > tmpfs running out of space. tmpfs can be in various locations. In my > case /dev/shm was used and it was far to small. I found that out using > lsof and grepping for my UML instances. > > hope this helps yes, this was the reason. Many thanks for your tip. I had a hugde swap space and therefor a big tmpfs in /tmp and /var/tmp but my /dev/shm was just 512MB and this is a joke for 8-12 UML instances. Maybe the 2.6.17.x kernel uses more /dev/shm(?); this problem never occours before. Now my shm is about 2G and the host and all UMLs are well alive :) Thanks and greetings, stefan |