On Thu, 2006-07-06 at 11:03 +0530, Anil Gupte wrote:
> Have you tried 20-30 per day? Everyday? That is the kind f stuff we have to
^^^^^^^^^^^^^ yipes!
> deal with here. Sometimes it is just fluctuation in the voltage and that is
> enough to reboot the system.
>
> Are you doing a standard install? Or do you have anything different in terms
> of
> partitions or file systems?
>
> Thanx,
> Anil Gupte
>
>
Our power is reliable. But sometimes a circuit breaker pops when the
air conditioner goes on. My wife (nearly) invariably resets the breaker
for the circuit that has the server on it. This does not fix the
problem of interest to her. So she resets it a few times until she
remembers which breaker will work. Meanwhile the server may have gone
through two or three partial boots in two or three minutes. This would
scramble RH 9 into a knot.
Debian stable (Sarge) with ext 3 has survived this at least half a dozen
times. It is a standard install. I have the home and backup directories
on hdb, and the rest of the system on hda (a smaller, older disk).
Webpages and the like are also on hdb, with soft links from hda if
Debian expects them in /var.
I think that the good recovery is due to ext3 and the way Debian handles
bringing up the interfaces. In my case it brigs up:
iptables for the firewall and internal proxy service (squid)
ppp to the dsl provider
a daemon to update the ipaddress for the no-ip.com nameservers
resolveconf to keep the provider nameserver names up to date and
coordinated with dhcpd
eth0 setup as the wired intranet
ath0 setup as master for the wireless intranet
Most of the time apache2 starts before the ipaddress has been propagated
by no-ip and it does not find the address of the virtual servers I have
specified. To recover from this I have to restart apache via ssh (this
box does not have a monitor).
So far no files have been lost, but this box does not do much io except
possibly for squid.
--
To UNSUBSCRIBE, email to [EMAIL PROTECTED]
with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
