[SSC Serv] SCC-Serv on x64

[Wayne Tucker]

On Sat, Jan 3, 2009 at 6:14 AM, Florian Forster <octo at ssc-serv.com> wrote:

> > 2.) The service crashes every now and then and has to be restarted.
> > The last error code was:
> > Faulting application SSCServ.exe, version 1.0.0.0, time stamp 0x4935b7af,
> > faulting module SSCServ.exe, version 1.0.0.0, time stamp 0x4935b7af,
> > exception code 0xc0000005, fault offset 0x00009ab7, process id 0x4f4,
> > application start time 0x01c96131a6601018.
> >
> > I'll keep monitoring this one and let you know if the exception codes
> > and fault offsets vary
>
> Eek, that's not good. Yeah, please let me know that. I'll see if I can
> find anything out with that information, but just the file offset is
> probably not going to help much :( What version did you use? It says
> 1.0.0.0 in the file because setting the version for a file it well
> hidden and I tend to forget to change that..
>
> > 3.) The CPU usage graph will sometimes spike way off the chart (e.g. 7
> > million % utilization)
>
> Sounds like a counter reset. This could be caused by automatic
> restarting of SSC Serv.
>
> > 4.) Occasionally, data will end up in weird places on the collectd
> > end.  For instance, there's a df/memory-cacched.rrd file with 1 sample
>
> Can you maybe create a PCAP file for a host that's having this problem?
> Bruno Prémont has recently written a Wireshark dissector for the
> collectd protocol, which is wonderful for problems like this one. Good
> arguments to tcpdump would be:
>  # tcpdump -i $IF -p -s 2000 -w sscserv.pcap udp port 25826 and host $IP
>

I upgraded to 1.2.2 last week.  So far it seems to be working flawlessly.

- Wayne
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.verplant.org/pipermail/ssc-serv/attachments/20090202/7d2155dc/attachment.htm