Only a couple of responses, but thus far it seems to fit.
I had two responses that this occured to others during Y2K testing.
Clock forward, clock back, and this can happen.
Well, I was skeptical because of the nature of this machine, but I
checked and sure enough, there's some Jan 1 dates in the log files. So
much for anyone telling me about testing on this box!! :-)
Thanks for the help!
Original posting follows...
> I've got some 2.6 servers that folks are developing/testing on. In the
> past couple of weeks, they've 3 times managed to do this on 3 different
> machines.
>
> Whatever they've done has triggered something that ties up every port on
> the machine...
>
> servera% netstat | wc -l
> 34137
>
> excerpt from netstat...
> [...]
> servera.37304 uberguy.32771 8760 0 8760 0 TIME_WAIT
> servera.37305 uberguy.32771 8760 0 8760 0 TIME_WAIT
> servera.37306 uberguy.32771 8760 0 8760 0 TIME_WAIT
> servera.37307 uberguy.32771 8760 0 8760 0 TIME_WAIT
> servera.37308 uberguy.32771 8760 0 8760 0 TIME_WAIT
> servera.37309 uberguy.32771 8760 0 8760 0 TIME_WAIT
> [...]
>
> servera is the machine in question.. uberguy is a NIS slave/DNS server.
>
> I ran lsof, and none of these ports have an associated process.
> However, they can't be used for any outgoing process..
>
> servera% telnet uberguy #(doesn't matter what I try to connect to)
> Trying 10.1.1.2...
> telnet: Unable to connect to remote host: Cannot assign requested address
>
> Ugh. I hate to just keep rebooting these machines when this occurs.
> I'm guessing that the kernel has become confused about some networking
> structure or something like that. Has anyone seen this before? Is
> there any specific cause? I'm running 2.6 with all the Recommended
> patches installed.
>
> Thanks for your info.
--
Darren Dunham ddunham@taos.com
Unix System Administrator Taos - The SysAdmin Company
Got some Dr. Pepper? Santa Clara, CA
< Please move on, ...nothing to see here, please disperse >
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:13:32 CDT