[AMBARI-12380] Ambari Metrics

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Invalid
• Affects Version/s: 2.1.0
• Component/s: ambari-metrics

Environment

Environment: The cluster is small with 3 nodes.
OS: RHEL 6.6
Platform: x series
Memory: 12GB on all 3 nodes
Ambari 2.1 version installed
All default configurations
Default master, slave, client assignments

• Epic Link: Ambari Metrics System
• Assignee: Unassigned
• Reporter: Chandana Mirashi
• Created: 10/Jul/15 21:17
• Updated: 28/Jan/16 19:16
• Resolved: 28/Jan/16 19:16

Description

Ambari GUI after few days is slow in responding to any of service stop and start commands . service timed out after 40 min. If AMS process was killed the service could be restarted quickly.

Ambari metrics processes takes up 27% of memory when seen using top command output on the metrics collector. The first process is the ams hbase master and 2nd process is ams timeline server.

top - 14:14:15 up 9 days, 2:34, 2 users, load average: 1.71, 1.70, 1.06
Tasks: 2 total, 0 running, 2 sleeping, 0 stopped, 0 zombie
Cpu(s): 13.5%us, 1.4%sy, 0.0%ni, 84.6%id, 0.4%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 11897780k total, 11760800k used, 136980k free, 9088k buffers
Swap: 4194300k total, 1020716k used, 3173584k free, 3178956k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
10794 ams 20 0 4073m 2.3g 7020 S 108.2 20.1 12379:06 java
10864 ams 20 0 8785m 1.0g 6776 S 0.0 8.8 150:04.19 java