I have the same problem of you, and i think i could have the purge function on local variable too in my taf, i 'll checked right away that to see if it's the cause of my problems.
Thanks Maggie
Le Lundi, 8 d�ce 2003, � 15:08 America/Montreal, MJPinckard a �crit :
I think I've found the problem. For memory management purposes (under T3.5) I was using local scope for my variables, and when I finished with a batch, would purge local scope.
In Witango 5 the next search action after a purge of local scope was failing. I found this because I tried changing the local scope to request scope (hoping that the problem might be related to the change in scope terminology), which didn't fix the problem but did make it obvious what the problem was. When I remove the <@purge scope="request"> I'm back in business with no crashes. I hope the purge scope isn't necessary and the built in memory management can cope with the variables hanging around (I have hundreds of variables)... I suppose I can purge each individually if need be... I have a batch of code doing that with no problems.
Hopefully all will proceed swimmingly from here...
Maggie
MJPinckard wrote:
I just tested that by changing the branch in the previous taf, so that rather than pointing back to the suspect taf where it has been dying, it points to a new taf. At the same point in execution (e.g. X actions after the branch) in the new taf the server crashed. I don't think it is the code of any particular taf, it is the process. How can I monitor what is going on in the process?
Maggie
Jeff Bohmer wrote:
To me, it just means your server crashed ;)
But now at least you're getting the proper messages in witangoevents.log and that's critical.
At this point, if I were in your shoes, I'd rebuild the TAF. Hopefully, it isn't huge. I know you've tried replacing actions, but that may not replace everything involved with the code being executed. Since the crash always happens with this TAF, it follows that replacing the entire TAF will either fix the problem or finally rule out the TAF as the culprit.
- Jeff
-- * * * * * * * * * * * * * * * * * * * * * Margaret Pinckard
* [EMAIL PROTECTED]
* Principal Research Associate
* End-use Forecasting Group
* E.O. Lawrence Berkeley National Laboratory
_______________________________________________________________________ _
TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf
________________________________________________________________________ TO UNSUBSCRIBE: Go to http://www.witango.com/maillist.taf
