Grouper Users - Open Discussion List

["Gettes, Michael" <>]

daemon_1    | /opt/grouper/grouper.apiBinary/bin/gsh: line 135:   114 Killed                  ${JAVA} -Xms$MEM_START -Xmx$MEM_MAX -Dgrouper.home="$GROUPER_HOME/" -Dfile.encoding=utf-8 $GSH_JVMARGS -classpath "${GROUPER_CP}" $GSH $*

From what I can tell, it is not running out of memory.  It was one of the things I was looking at… 

/mrg

On May 23, 2018, at 9:36 AM, Black, Carey M. <> wrote:

Michael,

Just a guess based on:  " By “killed” I mean the container generates an error message indicating “killed” and the entire docker container process disappears "

Maybe you are seeing a message like this? (not for mysqld, but for some docker process? Or the process in the container?)
"kernel: Killed process 9163, UID 27, (mysqld) total-vm:2457368kB, anon-rss:816780kB, file-rss:4kB"


It sounds like the OS is forcing the process( whole container?) down due to memory constraints/other resources. ( REF: "OOM Killer" )
Can you get the full error message? ( If my guess is correct you can find it in the OS that is hosting the container(s). )
/var/log/messages   ???

# or go fishing...
grep oom /var/log/*
grep total_vm /var/log/*


My general guess is that you need to give the hosting OS more ram.
Given that PSPNG, Loader, UI, (WS?), DB are all sharing that same RAM pool.... you might need to just not run some of that at the same time on your laptop. :^/


I have found that my loader can eat 4G of RAM for "large group jobs" ( 100k+ ). (I was actually seeing it die with "Heap size" errors until I expanded the ram for the process.) However, it may also be true that the RAM growth is due to not one job, but several "large" jobs running at the same time too.

HTH.

--
Carey Matthew


-----Original Message-----
From: <> On Behalf Of Gettes, Michael
Sent: Wednesday, May 23, 2018 9:05 AM
To:
Subject: [grouper-users] PSPNG and tier packaging

Good morning all,

Lately, I have been doing a lot of testing using the TIER Grouper package.  Part of the package includes a 389 directory server (great choice!) and I have configured the environment to publish groups to 389 using PSPNG.  I have getting more and more intense with the work as time goes on to be able to effectively demo Grouper on my laptop.  Over time, the “daemon” container is killed.  I haven’t been able to figure out what is going on or why.  I decided to remove the “provision_to” attributes from the folder hierarchy I was putting into 389 and noticed the daemon container wasn’t being killed any longer.  The daemon container essentially runs gsh -loader and that is all.  Turning PSPNG back on again, and the daemon container gets killed again.  By “killed” I mean the container generates an error message indicating “killed” and the entire docker container process disappears - the gsh -loader is no longer running.

I hope this helps point in the direction of a problem.  I would really like to be able to reliably run PSPNG.  I am including my PSPNG config below.  When PSPNG is running, it works great with this config.  If anyone is curious, givenName is being used because eduPerson is not yet part of the 389 config within the TIER Grouper package.

/mrg