|
From: Grant \(ProtectionNET\) <gr...@pr...> - 2003-12-22 07:50:47
|
I did see that in the docs...
The docs didn't say if that settings was for all restart conditions (of =
which in case of an error I would be happy for a 5 sec delay) or just =
Wrapper.restart().
I'll try setting the value to zero - what other effects does this =
setting have ?
Grant
----- Original Message -----=20
From: Sal Ingrilli=20
To: wra...@li...=20
Sent: Monday, December 22, 2003 2:55 PM
Subject: RE: [Wrapper-user] Wrapper restart delay
i can give you a reason:
assume that a runtime exception in your program is causing a restart =
() request as soon as the app boots-up.
this would generate an inifinite loop (start/restart) and your app =
would end up taking 100% of cpu cycles.
so the delay is a good thing.
however i agree with you that it should be your call & responsibility =
to make sure that doesn't happen.
so, have you tried reading the documentation & for example setting =
wrapper.restart.delay in wrapper.conf (depending on what wrapper version =
you're running)?
also, have you accounted for the time taken for
1. your application to stop
2. the time taken for the OS to stop the service
3. the time taken for the OS to start the service
4. your application to start
-----Original Message-----
From: wra...@li... =
[mailto:wra...@li...]On Behalf Of Grant =
(ProtectionNET)
Sent: Sunday, December 21, 2003 8:38 PM
To: wra...@li...
Subject: [Wrapper-user] Wrapper restart delay
Hi,
When my java app requests a restart via Wrapper.restart() there is a =
noticable 30 sec delay before the Wrapper restarts the VM.
I'm keen to reduce this delay to zero if possible. I'm only =
interested in reducing the delay when I request a restart - I don't =
understand why a delay exists when I request the restart. I would have =
thought this could be instant... no need to wait 30 sec is there ?
Is calling Wrapper.restart() the correct procedure to quit your app =
and instantly restart it ?
Regards,
Grant |