|
G'day,
We had a major power outage last night, and everything got rebooted,
including my LVS directors (RHEL3, Ultramonkey 3, mainly LVS-NAT).
Ok.
Everything came back up fine, but it seems that some Oracle clients
outside the cluster are timing out (i.e. connections seem to expire) to
the Oracle server on the inside of the cluster. I have set persistence
for these connections to 604800 seconds, but I forgot to script 'ipvsadm
--set 604800 0 0' to happen at boot time. I've just run this manually,
and I'll soon see whether it fixes the Oracle timeout issue.
I don't quite understand your problem because of the "Oracle clients
outside the cluster", are there any Oracle clients inside the cluster
that don't show this behaviour?
In the meantime, can anyone tell me what /proc values the 'ipvsadm
--set' command actually modifies? Something in /proc/sys/net/ipv4/vs?
Well, the current proc-fs entries are a read-only representation of what
could be set regarding state timeouts. The 'ipvsadm --set' command will set
a) The time we spend in TCP ESTABLISHED before transition
b) The time we spend in TCP FIN_WAIT before transition
c) The time we spend for UDP in general
for IPVS related connection entries.
Also, does anyone script an 'ipvsadm --set' command at boot time? If so,
where... /etc/rc.d/rc.local?
Can't comment on this, since I've never used it.
Regards,
Roberto Nibali, ratz
--
echo
'[q]sa[ln0=aln256%Pln256/snlbx]sb3135071790101768542287578439snlbxq' | dc
|
