[uml-user] Swap issues

[David C. <da...@da...>]

This isn't really a UML problem, but UML is the only thing which uses 
enough RAM to do it.

Everything runs fine for a while, then I get;

Sep 13 08:08:13 maeve kernel: swap_dup: Bad swap file entry de0e6020
Sep 13 08:08:13 maeve kernel: swap_free: Bad swap file entry de0e6020
Sep 13 08:08:33 maeve kernel: swap_free: Bad swap file entry ccb15020
Sep 13 08:09:17 maeve kernel: swap_free: Bad swap file entry de0e6020
Sep 13 08:09:46 maeve kernel: swap_free: Bad swap file entry c613c020
Sep 13 08:13:06 maeve kernel: swap_free: Bad swap file entry e4d2d020
Sep 13 08:13:59 maeve kernel: swap_free: Bad swap file entry f688a020

The box has two 2Gb swap filesystems, stored on a couple of LVM logical 
volumes, and /tmp is 3Gb. I'm guessing it's probably a bug or something 
in LVM, since the UML which kills it is probably thrashing the disk 
anyway, but I'm no further forward in tracking it down.

Anyone seen something similar?

David

-- 
David Coulson                                  http://davidcoulson.net/
d...@vi...                       http://journal.davidcoulson.net/