[Orca-users] New orca (and winallator) and data file cycling?

David Michaels dragon at raytheon.com
Wed Sep 8 13:38:20 PDT 2004


With the new orca (r380), I'm seeing this error every day when the 
respective server cycles to the next day's log file:

Orca: file '/npd/dmichael/orca-win/var/orca/winallator/Lenticular02/Lenticular02_Winallator_2004090609.tsv' was current and now is not.
  


Whenever this happens, Orca does not close out that file and move on to 
the new/current file, like it did (and still does) in 0.27.  While the 
above example specifically mentions winallator, I see the same behavior 
with orcallator-generated data files when they cycle.  It's like there 
used to be logic in there for advancing to the new file, and now that 
logic is replaced with an error and no advancing is performed.  This 
causes us problems, because the graphs stop getting updated for that 
host.  My work-around is to restart orca periodically.

Also, I've stopped using r380 for the orcallator file processing, since 
r380 doesn't seem to like the "volatile" keyword (even though I see code 
there to handle it).  The orcallator files are being processed by 0.27 
(which is working fine).

I have not had a chance to sift through the code to find the differences 
'tween 0.27 and r380 in this regard -- in the meantime, does anyone else 
have this problem, or is it just me?  Any ideas as to why it's happening?

--Dragon

-------------- next part --------------
An HTML attachment was scrubbed...
URL: </pipermail/orca-users/attachments/20040908/aa182e51/attachment.html>


More information about the Orca-users mailing list