Re: Illegal Seek Errors on label_result

From: Erin Charpentier (
Date: Thu Feb 01 2001 - 05:58:40 MST

Hi Dave,

I'm not sure if you've resolved your problem yet, but if you
haven't then this might be of help. I am running on Solaris,
and I came across this error today. In my case, anyway, I did
get the "no matching phases found" message that Duane suggested.
I couldn't get any results at all out of that log file. The
problem for me was that I sent a "kill -9" signal to the polyclt
process (from cron), rather than using Ctrl-C in the window it
was running in. I think a Ctrl-C (or maybe sending a TERM
signal instead?) is necessary. Hope this helps,


Duane Wessels wrote:
> On Thu, 25 Jan 2001, Dave Hecht wrote:
> > When trying to run label_results against a clt.log file, I
> > continually get '.....failed:probable cause; Illegal Seek.' I am
> > running on Linux. I have tried in vain to find the cause, or
> > documantation on this error - can anyone help me?
> Most likely there is another error before that. It probably
> says "no matching phases found".
> The report generator looks for certain phase statistics,
> probably the default is "top2". You can get this "no matching phases"
> error if:
> 1) your test doesn't have a phase named "top2". Maybe you
> changed the workload file.
> 2) polygraph died before reaching the end of the top2
> phase.
> 3) the log files are corrupted.
> Look for your reportgen.cfg file. You should see a line like:
> log_extractor_opts: --phases top2
> If you comment out that line report gen will print stats from
> all available phases.

This archive was generated by hypermail 2b29 : Tue Jul 10 2001 - 12:00:17 MDT