Thanks Matt for the pointer. But I don't see any thing mis-configured in
my setup. It was working absolutely fine until day before yesterday,
since I upgraded to the 7.4.4 and started using an alias name for the
Central manager. Nothing actually changed on the submit host, where
Schedd is crashing.
Any idea about any possible place I should look into?
Cheers,
Santanu
On 18/11/2010 12:56, Matthew Farrellee wrote:
You could try using a non-alias. I don't see how that'd crash the
Schedd though.
The "IP address of FALSE" is concerning, and the crash in
__nss_hostname_digits_dots. Maybe you have something really funny in
your configuration, say a non-IP (maybe FALSE) in a place where you
need an IP (w.x.y.z).
Best,
matt
On 11/18/2010 07:52 AM, Santanu Das wrote:
For the background info, I'm using alias name for the Central
Manager -
does it work?
Cheers,
Santanu
On 18/11/2010 12:40, Matthew Farrellee wrote:
Stack dump for process 2638 at timestamp 1290067960 (19 frames)
condor_schedd(dprintf_dump_stack+0xb3)[0x5c0e7a]
condor_schedd(linux_sig_coredump(int)+0x28)[0x5b1b68]
/lib64/tls/libpthread.so.0[0x3339b0c5b0]
/lib64/tls/libc.so.6(__nss_hostname_digits_dots+0x47)[0x33394d7a87]
/lib64/tls/libc.so.6(gethostbyname+0xac)[0x33394dab6c]
condor_schedd(condor_gethostbyname+0xa9)[0x5c5e8a]
condor_schedd(Scheduler::negotiate(int, Stream*)+0x3b8)[0x518d72]
...
and
11/18 08:12:19 (pid:2638) IPVERIFY: unable to resolve IP address of
FALSE
Something is probably strange with your network setup.
Best,
matt