Bug 88507 - Impossible to interrupt app server restart or stop
Impossible to interrupt app server restart or stop
Status: NEW
Product: serverplugins
Classification: Unclassified
Component: Sun Appserver 9
5.x
Sun All
: P3 (vote)
: TBD
Assigned To: Vince Kraemer
issues@serverplugins
:
Depends on:
Blocks:
  Show dependency treegraph
 
Reported: 2006-11-02 13:32 UTC by emcmanus
Modified: 2009-12-10 12:56 UTC (History)
0 users

See Also:
Issue Type: ENHANCEMENT
:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description emcmanus 2006-11-02 13:32:08 UTC
Using Sun Java System App Server 9 Update 1, if I change the JVM options so that
the server will fail when launched, and then I try to restart the running server
with right-click on the appropriate entry in the Runtime->Servers list, the IDE
will show a little animation in the bottom right-hand corner and a message
"Stopping [or Restarting] Sun Java System Application Server 9".  There is then
no way to interrupt the stopping process.  You have to wait for it to time out,
or else quit the IDE.

I understand that adding some explicit interrupt action could be problematic --
if you interrupt the stop or restart then the IDE might no longer know whether
the app server is running.  But it will be in that state after the timeout
anyway so there is nothing to be gained by forcing the user to wait for that
timeout.

While the restart is hung, a right click on the dialogue will bring up a menu
that includes "cancel process", but that item is greyed out.
Comment 1 Vince Kraemer 2008-04-25 20:05:20 UTC
please evaluate this against the v3 plugin...
Comment 2 _ pcw 2008-04-25 21:09:22 UTC
This request still applies vs. V3.  Such a timeout would still happen (currently 2 minutes, possibly could be shortened).

There are a couple of possible improvements:

1) We can and probably should monitor the server log stream and see if we can identify specific server death throes
where we can short circuit the timeout and kill the server automatically.  There are definitely cases that are detectable.

2) For cases not covered by #1, consider that the current implementation protects itself against synchronization issues
 by simply disallowing the user to affect the server state while the state is in flux (e.g. stopping/starting).

If we allow some form of "cancel start", we must ensure much more extensive synchronization correctness due to the
nature of the new states that will be allowed.  This is possible.  It may not be worth the effort.
Comment 3 Vince Kraemer 2009-12-10 12:56:42 UTC
no issue should be assigned to issues. redistributing the load


By use of this website, you agree to the NetBeans Policies and Terms of Use. © 2012, Oracle Corporation and/or its affiliates. Sponsored by Oracle logo