segmentation fault on power hardware

[mbruzek]

I am seeing a problem with the zend-server charm on the power hardware.

Here is the error in the Juju logs:

2014-09-25 16:41:46 INFO juju-log database:3: adding ZS to cluster with /usr/local/zend/bin/zs-manage server-add-to-cluster -n zend-server/0-10.0.3.31 -i 10.0.3.31 -o 10.0.3.128 -u jokaafeipuraere -p niveidoemaecooj -d ZendServer -r 60 -w 5 -s -N adminKey -K f13b874b8eac6aebb8f0244716f97dc02adf12bf10ced201771e755c4691923e 2014-09-25 16:44:30 INFO database-relation-changed NODE_ID = 1 2014-09-25 16:44:30 INFO database-relation-changed WEB_API_KEY = adminKey 2014-09-25 16:44:30 INFO database-relation-changed WEB_API_KEY_HASH = f13b874b8eac6aebb8f0244716f97dc02adf12bf10ced201771e755c4691923e 2014-09-25 16:44:30 INFO juju-log database:3: Adding vhost 10.0.3.31 2014-09-25 16:44:32 INFO database-relation-changed vhostInfo 2 10.0.3.31 80 false PendingRestart
2014-09-25 16:44:33 INFO database-relation-changed Segmentation fault (core dumped) 2014-09-25 16:44:33 ERROR juju.worker.uniter uniter.go:486 hook failed: exit status 139

The full log can be found here: http://paste.ubuntu.com/8427430/

I do not know what other logs are needed to diagnose this problem.

[afroyd]

Is it possible to give me a server in this state for debugging? Even better would be a way to run the charm in the power8 environment. So I can make sure I can reproduce it and it doesn't happen again after the fix.

[mbruzek]

I only have access to one power system that I am using for other work. I would be happy to collaborate with you via a Google Hangout or some other technology where I can share my screen with you. Otherwise I am trying to get access to the IBM server that is going to be used for the demo. I had to email Tim Robinson (timro@us.ibm.com) at IBM for VPN access.

[afroyd]

Thanks Matt, I'm using the iic machines for the demo and can reproduce the crash.