TREX Issue- "ERROR WHILE INDEXING" in some connectors

[sbsrajendran]

Hello Experts,

    We are facing an issue like "Error while indexing" in 10 connectors out of 32. We are facing this issue for a 3rd time and this is 2nd in a week. When we faced this issue last time we deleted all the connectors and recreated again. The indexing took nearly 7 days to complete. Now we don't want to do the same again i.e., deleting and recreating the connectors. Instead is there any possibility to resolve this issue?

    Connectors in error are:

• HRTMC AES Documents
• HRTMC AES Elements
• HRTMC AES Templates
• HRTMC Central Person
• HRTMC Functional Area
• HRTMC Job
• HRTMC Job Family
• HRTMC Org Unit
• HRTMC Position
• HRTMC Structural authority
• HRTMC Talent Group

    TREX server version we are using is 7.10.49.00

    PFA of Job logs for the connectors.

    Kindly help us to resolve this issue. Any help greatly appreciated and points will be awarded.

    Please let me know if your require any further information.

Thanks & regards,

Sakthi

[MartinHastik]

Hi Sakthi,

first of all 7 days for indexing is very much. Do you have so much objects in the orgmanagement?  Indexing takes for the most of my customers between 1 and 10 hours.(Sometimes below 1 hour)

There is a dependency between the connectors you mentioned above. So maybe in fact there is only on connector with an error. Please go to esh_cockpit and select the HRTMC_CENTRALPERSON and drill down into the physical Indexes. Check which index has an error. If it is e.g. HRTMC_PERSON you can reindex only this one. This is not so timeconsuming..

best regards

Martin

[sbsrajendran]

Hi Martin,

     Thanks for your reply.

     Yeah there is a dependency between the connectors but we did a full indexing by deleting and recreating all the connectors and for that it took 7days to complete.

     Is there any idea on for full indexing of above mentioned connectors how much time it will take approximately or how can we speed up the indexing?

Thanks & Regards,

Sakthi

[Former Member]

Hi Sakthi,

How do you run full re-indexing? You need to use ESH_COKPIT and do following steps:

1) Select all SOCs and delete those.

2) Create SOCs for your software component

3) Check that there are no TREX background jobs running in SM37 (RPTMC_CREATE_CHANGEPOINT_AUTH usually causes most of the troubles)

4) Schedule initial indexing in ESH_COCKPIT (here please schedule 2 objects with different schedule:

- USER_AUTHORITY - Daily schedule without real-time indexing flag (unflag Start immediately)

- HRTMC_REL_CP_P_209 - Same time as authority index, but with real-time indexing flag

Last point is quite important as authority tables are usually the biggest in ammount and change frequently by users (hiring, changing etc.).

Hope this helps.

KR,
V.

[sbsrajendran]

Dear Vladimir,

Again our TREX is not working. Planning to do Re-indexing again i.e.,by deleting & recreating the connectors. Here we need your suggestions to complete this activity without any error in indexing.

So pls share your view/experience on the same.

Thanks in advance.

Regards,

Sakthi

[Former Member]

Hi Sakthi,

As said you

You need to use ESH_COKPIT and do following steps:

1) Select all SOCs and delete those.

2) Create SOCs for your software component

3) Check that there are no TREX background jobs running in SM37 (RPTMC_CREATE_CHANGEPOINT_AUTH usually causes most of the troubles)

4) Schedule initial indexing in ESH_COCKPIT (here please schedule 2 objects with different schedule:

- USER_AUTHORITY - Daily schedule without real-time indexing flag (unflag Start immediately)

- HRTMC_REL_CP_P_209 - Same time as authority index, but with real-time indexing flag

V.

[Former Member]

Hello Sakti.

What his exactly the error raised on Trex system. Please have a look on the TREXIndexServer trace logs?

What is the hardware configuration of you TRex?

Have you think to update Trex to the latest Revision?

There is some possibilities to speed up TREx indexing? Yes there at least one complex scenario to speed up the indexing. But It requires at least 2 trex systems + a clone of production system recently refreshed.

In the past a complete indexing that past more of 7 days was reduced to  5 days and an unavailability of "only 12 hours for the last delta indexing. It includes lots of manual operations and two SAP Systems with the SAME SID...

regards

Julien

[S_Sriram]

Hi Sakthi

1. Are you doing the full or delta index? during the full index you have to bypass index search or lock the transaction in SAP system.

2. Once the index completed you can take the index object  backup using the Python script refer the SAP Note 975965 - TREX 7.0/7.1: Data backup (online) and recovery

BR

SS

[sbsrajendran]

Hi Sriram,

     Thanks for your reply.

•      We have done both full indexing and delta indexing. Could you pls explain this line in detail "during the full index you have to bypass index search or lock the transaction in SAP system".

•      What is the use of this data backup note whether we can use the TREX search when we are indexing if we implemented this note?

Thanks & Regards,

Sakthi

[S_Sriram]

Hi Sakthi

    We have done both full indexing and delta indexing. Could you pls explain this line in detail "during the full index you have to bypass index search or lock the transaction in SAP system".


     What is the use of this data backup note whether we can use the TREX search when we are indexing if we implemented this note?

Since you said that the full index takes 7 days, before deleting the connector take the full backup of business index objects using the python script and recreate the RFC connector restore the business index object from the backup.

BR

SS