Can not create MTE: no more space

[Former Member]

Hi All,

First of all "HAPPY NEW YEAR" to every one.

I am getting some alerts like "Can not create MTE: no more space" in my BW production system.

To resolve this SAP suggested to increase the value of parameter alert/MONI_SEGM_SIZE.

Current value : alert/MONI_SEGM_SIZE = 40000000(Dflt value)

Component version : SAP NetWeaver 2004s

So could you please suggest how much value we can increase,on which basis we have to increase and if increase more value what is impact?

Thanks & Regards,

Suresh Babu

[Former Member · Accepted solution]

You can remove unwanted MTE's and add the required ones.

[Former Member · Accepted solution]

Hi Raja,

Thanks for the reply,but we dont have any unwanted MTE to remove.Could you please tell me where all these alerts will save at OS level.

SAP Suggested to increase the value of parameter alert/MONI_SEGM_SIZE, i increase it but no use.

Paremer old value new value

alert/MONI_SEGM_SIZE 40000000 43232000

And after this again we are getting same alerts so SAP suggested to reduce the value of the DAYS_TO_KEEP_LOGS parameter which is defined in the methods RSPC_CCMS_AGENT and RSPC_CCMS_STARTUP, so i changed it to 3 days.

then we got 1% free space for Log Attributes, again after some time it is showing 0% available

and also throwing same alerts (Cannot create MTE

[4KGNP5XKDN4VGOO2SICE5RF32 01/10 04:31:29] (MTE class

[BW_MonAtt_Process_Chains]).

Could some body help me to resolve this.

Regards,

Suresh Babu

[Former Member · Accepted solution]

Hi All,

Now issue has been resolved with SAP Suggestion.

Soltuion:

-

---

To clear out the process chaini monitors, please do the following:

1. Open Transaction BWCCMS

Menu Item Extras

Activate Maintenance Fuctions.

2. Select a Process chain node from the monitoring tree as below.

BI Monitor

BIP

Process Chains

<Processchain name>

3. Select a log attribute(<Process chain log ID>) using the

check box and then delete the MTE(Shift+F2). A popup appears with four

options for deletion. Choose the option 4.This deletes all the log

attributes.

Now reset the monitoring status to warmup by performing the following

steps:

1. Execute Transaction RZ21.

2. Choose the radio button "Segment Overview" under Topology and then

click on Dispaly Overview. Now switch to change mode and select the

first row (Which shows the segment status ONLINE) , listed under the

tab "Local Segments(&)" and choose the button "Reset Segment in WARMUP

status".

The above mentioned steps clears the logs and from now on,the logs wouldbe kept for 7 days which is the default value.You can consider reducing

this value as well , by following the below steps.

1. Execute transaction RZ21. Select radio button 'Method definitions'

and choose button 'Display Overview' in section 'Methods'.

2. Double-click on method name RSPC_CCMS_AGENT. Goto tab 'Parameters'

and set a smaller value for DAYS_TO_KEEP_LOGS as per your needs.You

can maintain this value as 1. Please ensure that only the required

chains are monitored instead of all the chains.

After this there should be enough free MTE place and the

days_to_keep_logs parameter for the rspc_* agents will be considered

correctly.

Thanks & Regards,

Suresh Babu

[bill_cheng]

Hi B Suresh Babu,

we have the same situation.

The question is, if we implemented the solution in non production systems, we found the BWCCMS log attributes were deleted and never came back.

Is this harming the system CCMS monitoring?

Thanks and best regards,

cheng bill

[darryl_goveas3]

Hi Bill,

I noticed the same behaviour in our dev system.  Did you ever find out how to get the logs to start generating, or if was even an issue to worry about?

I did notice that the MTE for process chains is via an MTE class BW_MonCls_Process_Chains, and this class has no method assigned for Data Collection.  However, I do not seem to be able to update the data collection agent - it is locked to using a higher level MTE class (I can't find out which one this might be!).  I can see that MTE Class BW_MonAtt_Process_Chains is using that method for Data collection and method RSPC_CCMS_ANALYZE for Analysis.

Thanks!

Darryl

[Former Member]

Hi,

1. alert/MONI_SEGM_SIZE

The monitor tree elements (MTE/node) as well as the active Alerts of the CCMS monitor architecture are stored on every application server in the what is referred to as monitoring segment in the shared memory (Shm key 13 under UNIX by default in pool 10). The size of this segment (in bytes) is determined by the alert/MONI_SEGM_SIZE profile parameter.

The default value of this parameter is

Kernel 640: 40000000 (Unicode systems: 60000000)

Kernel 700: 40000000 (Unicode systems: 60000000)

The minimum value of the parameter is 2000000.

If alert/MONI_SEGM_SIZE is set to a smaller value, the CCMS monitor architecture is deactivated, and the application server in question can no longer be monitored in transaction RZ20.

Caution: We recommend that you do not set the value of the profile parameter alert/MONI_SEGM_SIZE to 0. This setting results in the deactivation of the CCMS monitor architecture and all other cyclical system programs.

As a result, background processing and logon load balancing no longer work (among other things).

2. If you want to increase the default value, please follow SAP note 731165 - RZ20: CCMS parameters for segment control. You need to calculate it based upon your system and requirements.