[cvsnt] Error with 2.0.51d ends with "results may not be corr ect"

[Lehman, Curtis]

Hi Tony,

I have to many users to know for sure what's going on. Part of the problem
is that since it's a runaway process that slowly uses up all the memory till
the system crashes I don't know for sure when the problem started, so I
can't tell which users caused it. Right now I have nothing in my historyinfo
file and my config file has comented out #LogHistory=TOFEWGCMAR. How can I
modify these files to get more of what cvs is doing logged for debug
purposes? (I have looked for documentation on these options, but  I haven't
found anything helpful.) Right now I have a history file and it seems to
show a little info, but not enough to know what's going on. A couple of the
lines are:

O4138d04a|clehman|LOCALHOST/*0|adit3104||adit3104
M4138d4dd|abueno|ALEX-BUENO-LINU|adit3104/openrg/pkg/main|1.3|run_ext_proc.c

How can I get some sort of timestamps and the full command line the user
entered?

Thanks,
Curtis Lehman

-----Original Message-----
From: Tony Hoyle [mailto:tmh at nodomain.org]
Sent: Friday, September 03, 2004 11:06 AM
To: cvsnt at cvsnt.org
Subject: Re: [cvsnt] Error with 2.0.51d ends with "results may not be corr
ect"

Lehman, Curtis wrote:
> on a windows box.  Could CVS be crashing because of the lockserver
> duplicating mixed-case locks, and causing the repository to fail?

No, I doubt it.  A mixed-case lock problem could result in the same file
being locked twice though which could cause data to be lost, or a
checkout to get the wrong revision.

I've no idea what could cause your problem - I've never seen it...
cvs.cvsnt.org has clients from all sorts of OSs hitting it and that
hasn't had the problem (although that's a Linux server too).

What were the clients doing when they caused the problem?

Tony
_______________________________________________
cvsnt mailing list
cvsnt at cvsnt.org
http://www.cvsnt.org/cgi-bin/mailman/listinfo/cvsnt