48 day uptime problem?

Torsten Duwe (duwe@ns.lst.de)
Wed, 28 Apr 1999 12:29:09 +0200 (MEST)

Rahul> Two of our machines running kernel 2.2.1 had networking problems
Rahul> after 48 days of uptime. Both of these are DEC alphas (miata),
Rahul> running what was Red Hat 5.0 but nearly everything has been
Rahul> upgraded.
[...]
Rahul> configuration and same kernel. In both cases it was after exactly
Rahul> 48 days of uptime. Since all other services were manually
Rahul> restarted I think it's a kernel problem.

Guessing wildly but educatedly, I'd say someone is counting timer ticks in a
32-Bit unsigned. The alpha's timer interrupts at 1024Hz AFAIK which would
result in said 48 days.

Rahul> Has this been seen before, and is it fixed in the newer kernels?

Haven't looked into it (see above), but I hope it will be fixed, otherwise
the i386 machines will have the same problem, even if only after over a year.
(Yes, linux is provably stable enough to keep up that long !)

Torsten

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/