|
From: Leif M. <lei...@ta...> - 2011-05-06 04:23:57
|
Gaurav, The restart in the log you sent me was definitely caused by an external source restarting the service. You can see how the Wrapper itself was stopped and then started again 3 minutes later. --- STATUS | wrapper | 2011/05/03 21:52:18 | <-- Wrapper Stopped DEBUG | wrapper | 2011/05/03 21:55:16 | Allocating a console for the service. DEBUG | wrapper | 2011/05/03 21:55:16 | Found console window. DEBUG | wrapper | 2011/05/03 21:55:16 | Wrapper console window hidden successfully. STATUS | wrapper | 2011/05/03 21:55:16 | --> Wrapper Started as Service --- Prior to the shutdown at 2011/05/03 21:52:18 however you could see that the Wrapper was running its 4th JVM. This means that the JVM had been restarted 3 times for some other reason. If you could send me the full log file that shows those restarts in a direct email I will take a look at what caused them. Those may be because of a timeout but I am unable to say from what you sent me. Cheers, Leif On Fri, May 6, 2011 at 1:04 PM, Gaurav Pgoel <gau...@tc...> wrote: > > Thank you for the replies. > Leif, you mean the service was restarted and we can ignore the fact that > the JVM was hung or not responding? > > There is one more observation that i want to share. We have seen this > similar behavior on one more client machine. This raises a doubt as to > whether both places did someone restart the service? > > Regards. > Gaurav Goel > Tata Consultancy Services > Mailto: gau...@tc... > Website: http://www.tcs.com > > ____________________________________________ > Experience certainty. IT Services > Business Solutions > Outsourcing > ____________________________________________ > > > From: Leif Mortenson <lei...@ta...> To: > wra...@li... Date: 05/05/2011 08:14 PM Subject: Re: > [Wrapper-user] Wrapper restarts JVM when no response from server > ------------------------------ > > > > Gaurav, > Thank you for the detailed message. It makes it easy to respond. > > In your case, the Wrapper is shutting down because the Windows Service > Manager has sent in a stop command: > > DEBUG | wrapperp | 2011/05/03 21:52:13 | read a packet PING : ping > DEBUG | wrapper | 2011/05/03 21:52:16 | SERVICE_CONTROL_STOP > > When the Wrapper stops, the Service Manager is starting it again > almost immediately. > > STATUS | wrapper | 2011/05/03 21:52:18 | <-- Wrapper Stopped > DEBUG | wrapper | 2011/05/03 21:55:16 | Allocating a console for the > service. > DEBUG | wrapper | 2011/05/03 21:55:16 | Found console window. > DEBUG | wrapper | 2011/05/03 21:55:16 | Wrapper console window > hidden successfully. > STATUS | wrapper | 2011/05/03 21:55:16 | --> Wrapper Started as Service > > This makes me think that someone pressed the Service Restart button in > the service control manager. But I am only guessing as we do not have > any way of telling what initiated such a restart. > > At the time that the Wrapper shuts down however, I do see that you are > on the 4th Java invocation since the Wrapper was launched. The log > fragment that you sent does not include any of the 3 restarts that > took place so I am unable to see what the circumstances were around > those restarts. Could you please locate one of them in your logs > and then send me a few minutes before the restart through a minute or > so after? > > I also see that you are using 3.5.4. The 3.5.8 version that we just > released contains some improvements which will prevent unneeded JVM > restarts on heavily IO loaded systems. I have not seen anything yet > to show that that is the problem however. > > Cheers, > Leif > > On Thu, May 5, 2011 at 7:48 PM, Gaurav Pgoel <gau...@tc...> wrote: > > > > Hi, > > > > I am using Wrapper 3.5.4. I have written a java code and running it as > java > > wrapper service on Windows NT machine. > > I have a code to uplaod files from client machine (on which service is > > installed) to server. > > Sometimes, it happens that the upload takes time say 5 - 10 minutes to > > upload a file. In the meantime, there is no response from server to the > > client. > > > > In this situation, I assume, that my wrapper service restarts JVM > thinking > > that the jvm is hung. > > > > Due to this, there is duplicate upload happening. To avoid this, i am > > thinking of increasing wrapper ping interval to say 300 seconds. What do > you > > think guys? > > > > Below is the wrapper.conf file contents: > <snip> > > > ******************************************************************************************* > > > > Also written below is the system logs; > > > ******************************************************************************************* > <snip> > > > ******************************************************************************************* > > > > > > Guys what your views on this? Is this really because Wrapper thinks that > the > > JVM is hung? Or is it because of some other reason? > > > > Thanks in advance to all. > > > > Regards, > > Gaurav Goel > > Tata Consultancy Services > > Mailto: gau...@tc... > > Website: http://www.tcs.com > > |