Re: [Wrapper-user] JVM restart with ping-timeout

[Leif M. <le...@ta...>]

Madis,
    The Wrapper sends PING packets to the JVM at an interval defined by the
wrapper.ping.interval property, every 5 seconds by default.  The JVM side of
Normally, the JVM immediately responds by sending a PING back to the
Wrapper process.
    A timeout occurs when ever more than the wrapper.ping.timeout has 
expired
without the Wrapper process receiving a PING packet from the JVM.

A) This is not a problem that I have ever had reported in the past.
B) Reviewing the source, I do not see any way at all that this could be 
happening
with the unmodified source.   Immediately after the "Received a 
packet..." message
is logged by the handleSocket() method, it proceeds to call the 
sendCommand()
method where a "Send a packet..." message should be logged.  That is not 
happening.

    I would love to help further, but it is not really possible for me 
to try to debug
modified versions of the Wrapper.   I have given you clues as to where 
the problem
likely is, but you are going to need to go in and debug what is happening.

    Cheers,
    Leif

madis wrote:
> I compile the service wrapper java source with my own project. Kinda 
> got that way because when I installed the service on XP sp2,  I could 
> not get it started. Everything was OK after I did a little restart but 
> I poked around a little before that (not much, just the properties and 
> dll resolving).
>
> Tryd with original source from version 3.2.3:
>
> Still doing 50% from full cycle:
> ....................
> INFO   | jvm 1    | 2007/03/14 16:18:13 | Received a packet PING : ping
> DEBUG  | wrapperp | 2007/03/14 16:18:17 | send a packet PING : ping
> INFO   | jvm 1    | 2007/03/14 16:18:17 | Received a packet PING : ping
> DEBUG  | wrapperp | 2007/03/14 16:18:21 | send a packet PING : ping
> INFO   | jvm 1    | 2007/03/14 16:18:21 | Received a packet PING : ping
> DEBUG  | wrapperp | 2007/03/14 16:18:25 | send a packet PING : ping
> ....................
>
> If the ping cycle is incomplete why does it take so many trys before 
> VM shuts down?
>
> Leif Mortenson wrote:
>> Madis,
>>     Looking at your logs, the JVM is receiving pings from the Wrapper 
>> process
>> correctly, but it does not appear to be sending the ping reply back to 
>> the Wrapper.
>>
>>     This is what a single ping cycle should look like:
>> ---
>> DEBUG  | wrapperp | send a packet PING : ping
>> INFO   | jvm 1    | Received a packet PING : ping
>> INFO   | jvm 1    | Send a packet PING : ping
>> DEBUG  | wrapperp | read a packet PING : ping
>> ---
>>
>>     But from your logs, you are only seeing the first two lines:
>> ---
>> DEBUG  | wrapperp | 2007/03/14 14:13:22 | send a packet PING : ping
>> INFO   | jvm 1    | 2007/03/14 14:13:22 | Received a packet PING : ping
>> DEBUG  | wrapperp | 2007/03/14 14:13:26 | send a packet PING : ping
>> INFO   | jvm 1    | 2007/03/14 14:13:26 | Received a packet PING : ping
>> ---
>>
>>     Reviewing the source, I can not think of any reason why this would 
>> happen
>> repeatedly.  There are event callbacks, but if the thread was getting 
>> stuck there
>> somehow, you would not see the second fourth line.
>>
>>     Are you using an unmodified version of the wrapper jar?  I ask 
>> because of
>> the lines like the following in the log:
>> ---
>> INFO   | jvm 1    | 2007/03/14 14:13:17 | Received a packet PROPERTIES : 
>> (Property Values)
>> INFO   | jvm 1    | 2007/03/14 14:13:17 | java props:set.WRAPPER_ARCH->x86
>> INFO   | jvm 1    | 2007/03/14 14:13:17 | java props:set.WRAPPER_BITS->32
>> INFO   | jvm 1    | 2007/03/14 14:13:17 | java 
>> props:set.WRAPPER_FILE_SEPARATOR->\
>> INFO   | jvm 1    | 2007/03/14 14:13:17 | java props:set.WRAPPER_OS->windows
>> ---
>>     From their timing in the log, they are being printed before your 
>> application has
>> been started and must be being displayed from within the internal
>> WrapperManager.readProperties method.  That is not output that was in 
>> the original
>> source.
>>
>> Cheers,
>> Leif
>>
>>
>> madis wrote:
>>   
>>> Hi
>>>
>>> Altough the system is practically idle I often get VM restarts beacuse 
>>> the ping request times out. I have set the timeout for 4 minutes with
>>> wrapper.ping.timeout=240
>>> wrapper.ping.interval=5
>>>
>>> Does anybody know how to fix this problem without disabling VM restarts?
>>>
>>> Madis
>>>
>>>
>>> Here is the logfile for an example restart:
>>>   
>>>     
>> <snip>
>>
>>
>> -------------------------------------------------------------------------
>> Take Surveys. Earn Cash. Influence the Future of IT
>> Join SourceForge.net's Techsay panel and you'll get the chance to share your
>> opinions on IT & business topics through brief surveys-and earn cash
>> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
>> _______________________________________________
>> Wrapper-user mailing list
>> Wra...@li...
>> https://lists.sourceforge.net/lists/listinfo/wrapper-user
>>   
>
> ------------------------------------------------------------------------
>
> -------------------------------------------------------------------------
> Take Surveys. Earn Cash. Influence the Future of IT
> Join SourceForge.net's Techsay panel and you'll get the chance to share your
> opinions on IT & business topics through brief surveys-and earn cash
> http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
> ------------------------------------------------------------------------
>
> _______________________________________________
> Wrapper-user mailing list
> Wra...@li...
> https://lists.sourceforge.net/lists/listinfo/wrapper-user
>