Hi

I am having a really random problem. I have a dual opteron proliant
dl145 system running Redhat ESv3. The problem is that for the last three
weeks I would have to reboot it about every 5 days because no new
network connections could be established to the machine. I will try to
ssh and I get a prompted for my password but after that it just hangs.
The same is true of telnet and ftp. Other things hangs like changing a
user password. Could I be running out of file descriptors or inodes?

When I do a strace on a hang process I get this ...
[root@scali rc2.d]# strace -p 19459
Process 19459 attached - interrupt to quit
select(4, [3], NULL, NULL, {6, 310000}) = 0 (Timeout)
select(4, [3], NULL, NULL, {10, 0})     = 0 (Timeout)
select(4, [3], NULL, NULL, {10, 0})     = 0 (Timeout)
select(4, [3], NULL, NULL, {10, 0})     = 0 (Timeout)
select(4, [3], NULL, NULL, {10, 0})     = 0 (Timeout)
select(4, [3], NULL, NULL, {10, 0}

[root@scali rc2.d]# netstat -ap | grep -i CONNECTED  | wc -l
142

unix  2      [ ]         STREAM     CONNECTED     2347840 19561/sshd:
root [p
unix  3      [ ]         STREAM     CONNECTED     2329799
19459/scaproxyd
unix  3      [ ]         STREAM     CONNECTED     2329798
19459/scaproxyd
unix  3      [ ]         STREAM     CONNECTED     2329797
19459/scaproxyd
.... about 81 entries of this which I can't kill.

Any tips would be greatly appreciated.

Thanks,
Tanya

[ Post a follow-up to this message ]