Hide Forgot
Description of problem: When trying to create jbosseap-6 app concurrently, it would fail with a core dumped error. Sometime this also happened when trying to create just one scalable jbosseap app. But it seems that this issue just could be reproduced on my testing env, which blocks the ose-2.0 performance testing. Version-Release number of selected component (if applicable): 2.0/2013-11-26.1 openshift-origin-cartridge-jbosseap-2.11.1-2.el6op.noarch How reproducible: almost always Steps to Reproduce: 1.Modify the ProxyTimeout to 600 in 000002_openshift_origin_broker_proxy.conf to avoid timeout error, create 5 scalable jbosseap-6 app concurrently. [root@broker ~]# for i in `seq 1 5`; do rhc app create app$i jbosseap -s -predhat& done [1] 10099 [2] 10100 [3] 10101 [4] 10102 [5] 10103 ... ... Your application 'app2' is now available. URL: http://app2-1111.stress.com/ SSH to: 52aade1134c48ce149000001.com Git remote: ssh://52aade1134c48ce149000001.com/~/git/app2.git/ Cloned to: /root/app2 Run 'rhc show-app app2' for more details about your app. Unable to complete the requested operation due to: The server node.stress.com that your application is running on failed to respond in time. This may be due to a system restart.. Reference ID: 5fec8922705c05cae33c9b3e393509a5 Unable to complete the requested operation due to: The server node.stress.com that your application is running on failed to respond in time. This may be due to a system restart.. Reference ID: a2051d7c640f37178fc53d6767ce05b1 Unable to complete the requested operation due to: The server node.stress.com that your application is running on failed to respond in time. This may be due to a system restart.. Reference ID: 4b4c97f989d29006c907011520dfdcfe Actual results: Errors could be seen in the mcollective.log ------ Failed to execute: 'control start' for /var/lib/openshift/52aaad7234c48c60d60000a1/jbosseap CLIENT_MESSAGE: Starting jbosseap cartridge CLIENT_MESSAGE: Found 127.1.246.1:8080 listening port CLIENT_MESSAGE: Found 127.1.246.1:9999 listening port CLIENT_ERROR: Killed (core dumped) ------) Expected results: Should create apps successsfully Additional info: The whole mcollective log of creating app is in the attachment.
Created attachment 837158 [details] mcollective.log
This issue could be reproduced on my performance testing env with puddle 2.1.z/2014-05-29.3. When trying to create a jbosseap-6 application, would get failed as following: ... Using jbosseap-6 (JBoss Enterprise Application Platform 6) for 'jbosseap' Application Options ------------------- Domain: 00 Cartridges: jbosseap-6 Gear Size: default Scaling: no Creating application 'app1' ... Starting jbosseap cartridge Found 127.10.197.129:8080 listening port Found 127.10.197.129:9999 listening port Failed to execute: 'control start' for /var/lib/openshift/53a007cb34c48ce9ee0000ce/jbosseap Terminated (core dumped) After making some changes in /etc/openshift/resource_limits.conf as BZ#1043414 said, this issue disappeared, jbosseap-6 app could always be created. limits_nproc=500 memory_limit_in_bytes=5368709120 # 5G memory_memsw_limit_in_bytes=5473566720 # 5G + 100M (100M swap)
*** Bug 1110162 has been marked as a duplicate of this bug. ***
(In reply to Gaoyun Pei from comment #4) > > After making some changes in /etc/openshift/resource_limits.conf as > BZ#1043414 said, this issue disappeared, jbosseap-6 app could always be > created. > limits_nproc=500 > memory_limit_in_bytes=5368709120 # 5G > memory_memsw_limit_in_bytes=5473566720 # 5G + 100M (100M swap) For the record, in nonconcurrent situation on OSE 2.1, the limits_nproc change was sufficient -- bug 1110162 comment 10.