TREX indexing not showing the optimized values in prd

[former_member182426]

For trex setup we have followed the blog http://scn.sap.com/community/erp/hcm/blog/2010/09/20/setting-up-netweaver-embedded-search-trex-for-s...

In DEV and QAS it works perfectly. showing the no.of optimized details count as marked in below screen.

But in PRD, it's showing empty details.

When I click on Job Log for Central Person it's shows some internal session terminated error check in ST22.

See attached log from ST22.

And I did some search and find the thread http://scn.sap.com/thread/2076997 which describes related to this kind of error.

As explained in this thread http://scn.sap.com/thread/2076997 is it mandatory to have separate TREX servers for E-Rec and NAKISA.

Currently we have only one TREX server box for PRD, and we are using for this NAKISA and E-Rec.

In DEV and QAS, trex we have installed on E-Rec box itself and used it for NAKISA also.

Only in PRD we did separate TREX box due to memory and load.

TREXLOGST22.HTM.zip

[former_member182426]

Any idea or solution for this error?

[Former Member]

Hi Shankar,

As i mentioned earlier SES and ESH using different technology and approach for searching method.

I already facing the same issue as your mentioned above.

I already implement based recommendation by SAP to separate box for ESH and SES in DEV, QAS and PROD.

Kindly if your log into OSS SAP.   

Regards

Mohamad Firhad Samuri

[former_member182426]

Hi Mohammad,

Thanks for your inputs.

But in my case where DEV and QAS are as below due to systems hardware availability less.

HR System separate + EREC System Separate and TREX also installed in this EREC system

NAKISA installed on HR System, and for TREX we connected to EREC(Where Trex is already installed in same EREC System). We did this for DEV and QAS it's working fine all indexes are generated with out any issues.

Only in PRD system we did separate HR System + EREC System + TREX System.

In this case we could not able to generate indexes for nakisa.

Do you think any other settings is reason for this?

Any how I will try to raise to OSS, in case any other inputs if they provide. Since we have already setup the systems and if we need to do separate for all landscapes it's little challenge again changing the systems landscape.

[Former Member]

Dear Shankar,

As advise from SAP is mandatory to have separate TREX servers for E-Rec and NAKISA(Talent Module).

Using TREXADMIN do you see have any red alert message for your production box.

Kindly if you can try delete manually table ESH_OM_LOGSYS from the system and do full indexing TREX using role SAP_ALL.

Regards

Mohamad Firhad Samuri

[former_member182426]

Hi Mohammad,

Yes, as you mentioned I am seeing one red alert message under Alert tab as you marked in your screen shot.

And in table ESH_OM_LOGSYS there is one entry.

Do you want me to delete this entry from table and re-generate the index ?

[Former Member]

Hi Shankar,

Yes, can you delete value table ESH_OM_LOGSYS and re-generate the TREX indexing.

Role user to indexing must have SAP_ALL or using BATCH_USER id.

Schedule job ESH_EX_FU_CPOINTER_REORG as per SAP recommendation, please refer to below link for detail maybe can be useful

http://wiki.scn.sap.com/wiki/display/TechTSG/%28ES%29+Performance+Problems+due+to+Heavy+Traffic+on+T...

http://wiki.scn.sap.com/wiki/display/BI/Real+Time+Indexing+-+Some+facts+-+Enterprise+Search

Regards

Mohamad Firhad Samuri

[former_member182426]

Hi Mohammad,

I just deleted all connectors this will delete entry from that table ESH_OM_LOGSYS also.

And then cleared all old spool first.

Then re-initiated the process, all connectors are created  and when I do initial indexing for few connectors, there is some error log as shown below.

For example see below error log for HRTMC JOB

30.12.2014 17:14:19 Job started
30.12.2014 17:14:19 Step 001 started 
(program ESH_IX_CRT_INDEX_OBJECT_TYPE, variant 2DC5AA4FBA02FC, user ID 
CONSULTANT)
30.12.2014 17:14:19 Start indexing for:
30.12.2014 17:14:19 
Object type name:HRTMC_JOB
30.12.2014 17:14:19 Connection 
GUID:002481A722421EE4A4849751791D19F3
30.12.2014 17:14:19 Logical 
system:PRDCLNT800
30.12.2014 17:14:29 Execute initial indexing
30.12.2014 
17:14:29 Populating index data failed for object type HRTMC_JOB in index 
ESH:PRD800~PRD800~HRTMC_JOB~OBJID
30.12.2014 17:14:29 Depopulating in 
dependent nodes unsuccessful for index 
ESH:PRD800~PRD800~HRTMC_JOB~OBJID
30.12.2014 17:14:29 Error executing DFD 
Query: Deletion from dependent node CAREER_TYPE failed
30.12.2014 17:14:29 
Search was unsuccessful
30.12.2014 17:14:29 TREX call terminated with 
technical error
30.12.2014 17:14:29 Remote communication failure with partner 
http://smctrx01:30003/searchCellTable
30.12.2014 17:14:29 Indexing of 
application data was successful
30.12.2014 17:14:29 Search object connector 
PRD800~HRTMC_JOB~: New status: Active
30.12.2014 17:14:29 Job finished

From above error we have seen some Remote communication failure, but when I do RFC test it's working fine from HR system

Remote communication failure with partner 

But few of them it's created fine.

[former_member182426]

Any inputs related to this error?

[Former Member]

Hi Shankar,

Can you check on the table se16n ESH_CO_CPOINTER.Delete all the value inside the table and do full indexing.

Could you try this approach for full indexing:-

Step 1 - Delete all connector using tcode se38

Step 2:Create connector using this link http://nakisa.aiapublic.my/?p=258

Regards

Mohamad Firhad Samuri

[former_member182426]

Hi Mohammad,

Thanks for your reply. As you suggested

Step 1 - I already did this way also.

Step2  - If you see my initial post we have followed the same steps for indexing from below blog.

http://scn.sap.com/community/erp/hcm/blog/2010/09/20/setting-up-netweaver-embedded-search-trex-for-s...

It's looks like something different issue related to basis, like network host.

[former_member217429]

Hi ,
the message in the Job log "

Remote communication failure with partner http://smctrx01:30003/searchCellTable"
means that the ICM connection between ESH/ABAP and TREX was broken . Unfortunately from
the log it's not possible to identify the exact reason of the problem. Do

you have any corresponded entries in the ICM (ta. SMICM) or TREXDaemon trace , f.e.
Connection TimeOut (icm log) , restart of IndexServer (TREXDaemon trace)?
What's the TREX version you use?

Best regards,
Mikhail

[Former Member]

Hi,

have you scheduled an reorganization job for spool ? The report is RSPO1041.

Daniele

[former_member182426]

Hi Daniele,

Thanks for your inputs.

As of now in PRD system is not schedule this report in background. Is this report  RSPO1041 mandatory to schedule in background job in PRD system or can we run it manually when ever we required.

If we are running it manually or scheduling it background can we select the standard SAP variant SAP&001 and execute it safely.

Please advise.

[Former Member]

Hi Shankar,

this report is recommended as an job reorganization, see note 16083.

By default all sap ABAP system has that jobs scheduled.

For test you can run once and you can create a variant different of the standard ones if you have some particular spool that you want to maintain for a long time; the report RSPO1041 has a great possibilities to choice the retention time and more; for example you can exclude something and so on;

othewise the variant SAP&001 is good, but pay attention, you have to specified the correct sap client, sometimes that variant is present only on 001 client but not in the production client.

Anyway I suggest to find the correct strategies for retention times of spool and schedule that report daily.

Daniele

[Former Member]

Dear Shankar,

It may be due to spool overflow, there may be so many reasons. And have you tried indexing once through your ses_admin and do a full indexing once and see if you find any error.

Regards,

Venkat D

[former_member182426]

Dear Venkat,

Thanks for your prompt response. But SES_ADMIN is for e-rec system right.

Even we have not done this in DEV and QAS HR systems. It's working fine.

Same process we are fowling in PRD and facing this issue.

We are at EHP5 and using NAKISA 401_702_RF 0000 - Nakisa STVN Add-on RF

Regards,

Shankar.

[Former Member]

Dear Shankar,

Good day to you.

eRecruitment Search Engine Service (SES) and Talent Management (ESH) using different

approach for search.

SES and Embedded Search can not shared the same box because the technology and

configuration different.

Kindly check with basis to install the SES and Embedded Search for different box.

Regards

Mohamad Firhad Samuri

[former_member182426]

Dear Mohamad,

Thanks for your reply.

As per your information, TREX needs to be installed separately for NAKISA and we can't use the existing TREX server which we are using for E-Rec for this ? Please correct me.

But, in DEV and QAS we have the system landscape as below

DEV

-> HR System one box (NAKISA installed)

-> E-REC system & TREX is in one box

-> EP one box (Connected to HR and E-REC systems)

Same landscape exist for QAS also.

Here in both cases we have created on RFC for NAKISA TREX which is installed in E-Rec box.

It's working fine with out any issues.

In PRD system, due to disk space and performance, we have separated the E-Rec and TREX as separate boxes. Now my dought is when it's working in DEV/QAS why it's not working for PRD when we have separate boxes for all.

I tried to find any OSS Note related to this which says TREX should be separate for NAKISA. Please let me know if you have any notes.

Regards,

Shankar.

[Former Member]

Hi,

Have you try running ESH_ADM_INDEX_ALL_SC and delete/create all your connector?

Starting that the spool overflow is genuine have you cleanup your spool (SPAD - Delete Old Spool Requests)?

Best regards,

Francois Valade

[Former Member]

Hi Shankar,

Kindly check below SAP Note.System landscape for Prd, Qas and Dev must be the same architecture.

TREX Master (talent) and TREX Master (E-recruitment) should be different server.

Search Engine Service(SES)

SAP Note 1254901 - SES for custome project on TREX search

Embeded Search(ESH)

SAP Note 1249465 - TREX 7.1:Installing TREX for Embedded Search

SAP Note 1628169 - TREX 7.10 for Embedded Search - Landscape restriction

Regards

Mohamad Firhad Samuri

[former_member182426]

Hi Mohamad,

As I explained in my previous post, in DEV and QAS systems there is no issue even when I am having TREX in one system with in E-Rec system.

Only for PRD system, we made separate system alone TREX and trying to use that for both (HR and E-Rec). Now my query is, when it's working in DEV and QAS (with in one system having E-Rec and TREX) , in PRD separate system it's not working.

Regards,

Shankar.