SERVER REBOOTING

[BINUKUTTAN_1]

HAI ALL,
Iam facing a serious issue of server rebooting in a critical server in a bank. Tjis happened after the smps has burnt in that server. It is a D220, hpux 10.2 server.It has sybase appl. The memory was 64 MB.

The server main board, processor card and smps was changed. In shutdown log it is giving Swap Device Error.
I increased Device swap space. still it persisted. but when I increased the server memory to 128 MB the rebooting suddenly stopped. But it is a temp . solution becuase the memory is of another cust.

pls give solutions on this isssue..

thanking you all BINUKUTTAN

[A. Clay Stephenson]

Hi,

I think your problem is not the SIZE of the swap but rather a bad (or at least bad sectors) on the swap disk itself. When you increased memory, the box no longer had to swap
and the problem seemed to go away. As soon as load increases and swapping starts the problem will reappear. If this is a critical server, then I would install Mirror/UX and mirror my boot/swap disks so that a disk problem would not be a serious issue.
This box probably survived burn-in because the tests dis not induce much swapping.

Hope this helps, Clay

[Stefan Farrelly]

You can check if it is the swap area which is reporting disk problems by doing a dd on it;

dd if=/dev/vg00/rlvol2 of=/dev/null bs=64k

This is presuming a standard install where primary swap is lvol2. Obviously, do this out of hours in case it does cause a crash. If it does, or reports I/O error, then you have a filesystem block problem as Clay suggests, which will involve rebuilding your system as the only real way to fix a bad block is mediainit (low level format) the disk. If dd report X blocks in+out then its fine, you have another problem. Ive seen Sybase cause spirrious crashes before, especially with so little memory (64 MB!).