Actions from Monitoring (view sysout, hold, free, etc.) are very slow after the installation of Control-M Workload Archiving

Knowledge Article

Article Number

000224181

Old Article Number

000118770

Article Type

Solutions to a Product Problem

Title

Actions from Monitoring (view sysout, hold, free, etc.) are very slow after the installation of Control-M Workload Archiving

Summary

Actions from Monitoring (view sysout, hold, free, etc.) are very slow after the installation of Control-M Workload Archiving

Product

Control-M Workload Archiving

Component

Control-M Workload Archiving

Applies to

Control-M Workload Archiving 9.0.00,9.0.18,9.0.19

Problem

After a new installation of Workload Archiving or if Archiving was stopped for several days when started, it may take some time to catch up. When this happens, it is possible that when performing actions from the Workload Automation GUI Monitoring domain for the jobs (view output, hold, free, etc.) that these actions be delayed up to several minutes or even fails with such message in the Gateway logs :
===> ERROR: EM out link buffer overflows. CommEMBufferSize = 100000000 Actual = 100001292

Cause

potential performance problems on the M/Server (or Control-M for z/OS)

Solution

In order to lower the impact for the connected users working with jobs from the problematic datacenter, there are several options:

1- Decrease the “number of collection threads” parameter in Workload Archiving by updating the {EM_HOME}/archive/config/SystemParameters.xml. This value can temporarily be decreased as low as 2, and then increased later by updating the following session and changing the Value. The Archiving server will need to be recycled for this change to take effect.
---
<Parameter>
<Name>collection.threadpool.size</Name>
<Value>12</Value>
<SelectedType>Integer</SelectedType>
<Description>number of parallel collections (per second).</Description>
</Parameter>
----

2- Increase the gateway buffer in the Control-M/Enterprise Manager System Parameters in the CCM by increasing the value of CommEMBufferSize from 10000000 to 100000000. The EM Gateway will need to be recycled for this to take effect.

3- Increase the maximum number of CS processes in M/Server in the ctm_menu -> 4 - Parameter Customization -> 2 - Advanced Communication and Operational Parameters -> 3) Maximum server processes (CS). A value of 10-15 is generally acceptable. The M/Server will need to be recycled for this to take effect.

4- Disable the collection of old log and sysout (introduced in 9.0.00.411). To make the change, add the following into the systemparameters.xml file.
Example of 2 days:

---
<Parameter>
                <Name>num.days.log.keep.in.server</Name>
                <Value>2</Value>
                <SelectedType>Integer</SelectedType>
                <Description>job log will not be collected if more than num.days.log.keep.in.server days passed from job execution </Description>
            </Parameter>
<Parameter>
<Name>num.days.output.keep.in.server</Name>
                <Value>2</Value>
                <SelectedType>Integer</SelectedType>
<Description>job output will not be collected if more than num.days.log.keep.in.server days passed from job execution </Description>
            </Parameter>
---
The Archiving server will need to be recycled for this change to take effect.

5- Disable the collection of cyclic job logs in the systemparameters.xml following the format in #4 above. The parameter names and valid values are:

DISABLE.CYCLIC.LOG.COLLECTION
Values: IF_ENDED_OK, IF_ENDED_NOTOK, ALWAYS, NEVER (default is NEVER)

COLLECT.DUMMY.JOBS
Values: TRUE, FALSE (default is TRUE)

The Archiving server will need to be recycled for this change to take effect.

Attachment(s):