----------------
> In article
>> ok. time is becoming critical and i will need to reboot soon. this is a
>> mass invite to diagnose a very hard to uncover TCP bug.
> Looks like you ran out of free memory. Could you check with "free" if you
> haven't already rebooted? Also try pressing Shift-ScrollLock on the console
> and send the resulting output.
the machine has 192megs of ram, 80 megs were free, i was an hour and a
half flight time from the console.
> SUID bit lost from the ssh executable? Ssh should be SUID root.
ssh was suid.
> Works fine for me here. Maybe you need to increase
> /proc/sys/net/ipv4/ip_local_port_range?
changing 1024/4099 to 1024/16384 in this file magically fixed it.
================
factoids:
lots of free ram
lots of free cpu cyles
lots of free disk space
lots of free sockets (100 odd sockets in use)
/proc memory looks very similar to other machines. i.e. no suspicions
in short, nothing was apparent. it was simply broken. users couldn't do
-anything- wrt opening a socket. root could open a socket only if it was
bound under 1023. with netstat/proc reporting just a bit over one hundred
sockets in use, (4999-1024) should be leaving quite a bit more room. the
machine wasn't loaded, cpu was zilch, disk zilch, network zilch. inbound
connections worked just dandy. local ports gave eagain/eaccess.
i'd toss out the idea of a magic number being hit in the local port code.
any takers on this?
-d
-- Look, Windows 98 Buy, lemmings, buy! MCSE, Must Consult Someone Experienced __ (c) 1998 David Ford. Redistribution via the Microsoft Network is prohibited \/ for linux-kernel: please read linux/Documentation/* before posting problems
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/