On Fri, Mar 28, 2008 at 02:12:49PM -0000, Ward, Martin wrote:
Hi All,
I am running Hobbit 4.2.0 on a number of Sun servers and recently the
hobbitd_client daemon has been crashing and leaving core files around.
[snip]
#7 <signal handler called>
#8 0x08063ef9 in get_ostype (osname=0x806c28d "") at misc.c:43
#9 0x0805b320 in main (argc=1, argv=0x80464ac) at hobbitd_client.c:1766
[snip]
The really bizarre thing is that it only crashes when one particular
machine connects to it and sends data.
Weird, indeed. It looks like the client message from that particular
host is garbled in some way - specifically, the "osname" argument should
never be blank, it should be "sunos" or some other operating system
name.
Could you have a look at this client host ? Specifically, the client
data that it sends to Hobbit are in the
~hobbit/client/tmp/msg.HOSTNAME.txt file over on the client - it
would be interesting to know what the first few lines of that file
contain. It *should* read something like this:
client my,hostname,com.sunos sunos
[date]
Sat Mar 29 07:40:29 MET 2008
[uname]
SunOS my.hostname.com 5.8 Generic_117350-46 sun4u sparc SUNW,UltraAX-i2
[uptime]
7:40am up 197 day(s), 22:10, 0 users, load average: 0.28, 0.16, 0.14
Regards,
Henrik