Lower amounts (anything above 1500 bytes) would cause varying problems..
first was just a memory leak, then it was a spontanous reboot, then at
the 64k+20 number, it went totally bizerk and totally locked up our
network - the passive hub we use here was reporting 100% throughput 100%
collisions (that was the night I found out it'd beep when the network
went haywire)... a reboot fixed all the problems above.
Strange concidence that the 64k+20 was VERY interesting to watch on an
old linux kernel (2.0.0 is choice).... It'd cause the box to die, and
in the process wish it was dead.
JS
Mark Morley wrote:
>
> The other day my pager started going off, telling me that one of our
> PM2's was down. Sure enough, I couldn't ping it, telnet to it, etc.
> We rebooted it and all was fine for 5 minutes, then it went dead
> again. Then all the other PM2's on our ethernet stopped responding
> as well. No error messages on the consoles, they just locked up.
> We'd reboot them, they'd be fine, then 5 minutes later they'd do it again.
> It finally stopped when we removed the PM2 that first went down. After
> that, no more problems.
>
> Then we noticed that one of our servers was reporting an IP conflict
> with a certain ethernet address. Sure enough, it was the PM2 that
> we had removed from the network. We hooked a console up to it and
> it showed no abnormal settings, it certainly wasn't in conflict
> with any other IP numbers on our network. It's been running fine
> for months and there were no recent changes made to it. Yet every time
> we put it on the ethernet it starts killing all the PM2's and various
> servers report duplicate IP numbers originating from this device.
>
> This happened once before, exactly the same way, except with a different
> PM2. At that time we had just upgraded it to the latest ComOS and
> rebooted it - wham, same thing as described above.
>
> When this happens it doesn't affect any of the PM3's on the same network,
> nor does it seem to affect any other devices (except for the console
> messages on the UNIX boxes about duplicate IP's).
>
> Does this ring any bells for anyone? Any idea what's going on? Could
> the ethernet port on these two units have gone south?
>
> It's quickly becoming a moot point as we are almost finished replacing
> all PM2's with PM3's, but I'm still curious...
>
> Mark
> -
> To unsubscribe, email 'majordomo@livingston.com' with
> 'unsubscribe portmaster-users' in the body of the message.
-
To unsubscribe, email 'majordomo@livingston.com' with
'unsubscribe portmaster-users' in the body of the message.