I have no real suggestions there but I can testify the
same kind of behaviour with the network setting during
the bootup phase as well. It looks like it is coming
from the fact that it tries to access DNS somehow before realizing the network is not
up yet so it gives up with DNS. (well, this is just a feeling
I have about what is happenning).
There are a couple of pbs like that this the redhat boot up.
Like setting up an alias for your eht0 when using DHCP for the
primary IP. If I use the linuxconf stuff to do this, it will try to
setup the alias during the bootup which will always fail (no asking me
why), but if I start aliasing manually in the rc.local everything is fine.
Philippe
"patrick" <[EMAIL PROTECTED]> writes:
> I don't understand this. I am running RH 6.1 and
> have pump-0.7.2-2 installed. I have been running
> setiathome for months and months nonstop as well.
> It usually sits at 80% or higher in CPU usage (with
> nice 19) if I am not really doing much. For the
> last two days, I have seen that it is sitting at
> about 3% CPU usage and no higher while pump is
> using 45% or higher of the CPU. Why in the world
> would pump need to use that much of the CPU for
> a prolonged period? It should only need to run
> for a few seconds at bootup to negotiate/get an
> IP address assigned to my box (DHCP with ADSL
> thru USWest).
> Can anyone enlighten me as to why pump would need
> to run for such a long time, or even if that is normal,
> why does it use so much of the CPU for such a long
> time? What do I need to do to fix this?
> The situation is that I used to have my box
> setup to start my eth0 at bootup. I have a 3Com
> EtherlinkXL PCI card that works nicely. Letting
> pump run at bootup after setting it up to do so
> via inetcfg led to a LOOOOONG wait when it came to
> eth0 initialization. Bootup would stop for a long
> time but then FINALLY continue with an OK for
> the eth0 initialization. This was intolerable so I
> unchecked the start at bootup option in inetcfg
> and opted to place /sbin/pump & into my
> /etc/rc.d/rc.local file outside the if statement.
> This allowed bootup to proceed quickly, with no
> inordinate delay while waiting for an IP to be
> successfully assigned to my box AND it would still
> allow me to immediately start up netscape or telnet,
> etc, after logging in. Whatever delay was occuring
> ceased to affect my use of the system upon doing
> this.
> Would my doing this somehow have led to this
> periodic OVERUSE of the CPU by pump? I cannot tolerate
> the incredibly long delay at bootup caused by pump
> if I select start at bootup for eth0 but it is also
> intolerable for pump to use so much CPU for such a
> long period (no activity lights flashing on my Cisco
> 675 router while pump is running thusly either) and
> deprive setiathome of CPU cycles.
> I am using a 2.2.13 kernel with my ethernet's driver
> compiled in - no as a module. The driver is the 3c59x
> driver.
>
> Any suggestions, aid, or explanations would be greatly
> appreciated.
>
> patrick
>
>
> --
> To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe"
> as the Subject.
--
To unsubscribe: mail [EMAIL PROTECTED] with "unsubscribe"
as the Subject.
