cloudstack-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Saurav Lahiri" <saurav.lah...@sungard.com>
Subject Re: Review Request 20123: Fix for cloudstack-6328 to Prevent console proxy support scripts from spawning multiple java processes
Date Mon, 28 Apr 2014 14:27:14 GMT


> On April 25, 2014, 10:10 a.m., Santhosh Edukulla wrote:
> > systemvm/scripts/run.sh, line 44
> > <https://reviews.apache.org/r/20123/diff/3/?file=558063#file558063line44>
> >
> >     Just to know, is _run.sh called only from here or has other references as well?
If yes, then this logic does nt solve purpose i believe. If it is only here, then it seems
ok. 
> >     
> >     AS well, i hope it is tested for various occurrences of java processes and the
sequence is working fine?
> >

Yes that's correct _run.sh is only invoked from run.sh and spawns the java process for systemvms.


> On April 25, 2014, 10:10 a.m., Santhosh Edukulla wrote:
> > systemvm/scripts/utils.sh, line 18
> > <https://reviews.apache.org/r/20123/diff/3/?file=558064#file558064line18>
> >
> >     is one time check for lock ok? Just use a timeout for acquiring lock, this way
it will atleast wait that many units to get a lock otherwise exit.

I think it should be a good idea to not try to acquire a lock because the first run.sh  which
has acquired the lock would already attempt to start the java process and retries by the second
run.sh process which has initially failed would not perhaps not be correct.


- Saurav


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
https://reviews.apache.org/r/20123/#review41457
-----------------------------------------------------------


On April 23, 2014, 9:20 a.m., Saurav Lahiri wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> https://reviews.apache.org/r/20123/
> -----------------------------------------------------------
> 
> (Updated April 23, 2014, 9:20 a.m.)
> 
> 
> Review request for cloudstack, Jayapal Reddy, Rajani Karuturi, Rajesh Battala, and Santhosh
Edukulla.
> 
> 
> Repository: cloudstack-git
> 
> 
> Description
> -------
> 
> With multiple java processes writing to the same logfile, each is not aware of the log4j's
internal counter state, this needs to be prevented. So before starting new java process via
the _run.sh , a check is made to ensure that there are no existing java processes running.
This will prevent multiple java process writing to the same log file namely cloud.out. 
> 
> 
> Diffs
> -----
> 
>   systemvm/patches/debian/config/etc/init.d/cloud 83853bc 
>   systemvm/scripts/run.sh 146d96f 
>   systemvm/scripts/utils.sh PRE-CREATION 
> 
> Diff: https://reviews.apache.org/r/20123/diff/
> 
> 
> Testing
> -------
> 
> Tested the changes with console proxy vm and secondary storage vm. They start and stop
as expected.
> 
> 
> Thanks,
> 
> Saurav Lahiri
> 
>


Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message