VMware Cloud Community
TheChristoph
Contributor
Contributor

Performance-Problem "Method ran a long time"

Hello again,

my problem is that my HQ-Server is really slow because of its nearly constant CPU-Performance of 100%.

I've got 44 Agents in my Hyperic-Observation, the Server on a single machine which not really fits the requirements (Windows2k SP4, 2x Intel Pentium 3 at 1133MHz, 4GB RAM). The Hyperic Version I use is 3.2.4.


I have seen in the log-files that 2 processes are constantly working and throw the warning "Method ran a long time":

Class: org.hyperic.hq.bizapp.shared.LatherBossLocal
Method: dispatch
RunTime: 2621292

Class: org.hyperic.hq.appdef.shared.AgentManagerLocal
Method: getEntitiesForAgent
RunTime: 920929

If the process has ended it starts again, and again, and again... (sometimes I have seen up to 4 of those big JBossLatherServlet - Stack Traces at Administration/HQ Health/Print "HQ Process Information")
...so the processors of my HQ-Server are constantly on 100%.

My questions are:
- What's the use of theses processes?
- Is it correct that they start and start again after finishing or is this a problem I may fix?
Reply
0 Kudos
2 Replies
excowboy
Virtuoso
Virtuoso

Hi,

Lather is the communication protocol for Agent and HQ Server communication. I think the server can not collect all metrics from the Agent due to the high load.
You will need to either buy a new machine, decrease the number of collected metrics or increase the collection interval for a lot of metrics.

hth,
Mirko
Reply
0 Kudos
TheChristoph
Contributor
Contributor

OK, I will decrease some metrics in time, maybe that could help.

Danke Mirko!


If I want to update the Hyperic-Server in the future (to a higher Version of HQ 3 or to HQ 4), would it rather be negative or positive for the performance on this machine?
Reply
0 Kudos