TREX Index Problem:

[Former Member]

Hello Everyone,

I have a problem with my TREX indexing. Please read below & advise:

1. Created a new index in index administration.

2. Selected crawler as standard, search as TREX Search and assigned a datasource as /documents/Mydocuments

3. Also entered a system ID for the Indexmanagement Taskqueue Reader Task.

When I look at the Index Monitor, I can see that the index has a green light with Active status,

In Process:23, Indexed:0

When I look at the queue:

Queue status: Idle

To Be Synchronized: 23 - It just waits in this state.

I tried flushing the queues and re-starting the services, but to no avail...

Any thoughts on what needs to be done to start the indexing process and fix my search?

Thanks and Regards,

Phani - SAP Basis Expert.

[former_member195891 · Accepted solution]

Hi Phani,

After flushing the queues (please without restarting of any services) what happens to the documents? They should all go to the "right" till the status ok (=synchronized) is reached.

In the standard the queue parameter Schedule Time (see TREX Monitor -> Edit Queue Parameters) is set to 30min (ALL-0:30) which means that the documents stay in status To Be Synchronized for at least 30min. For test purposes e.g. it might be helpful to set the parameter to ALL-0:01. In this case the documents will be synchronized immediately.

Kind regards,

Roland

[Former Member · Accepted solution]

Hello Mittal & Roland,

Thanks for taking the time to respond.

When I flush the queues from TREX Monitor -> Select the index -> Change queue status -> Select the Index ID and Flush - The queue status still remains as Idle.

The application log messages are:

Delta indexing failed. AbstractTrexIndex: The index server (for the specified indexes) cannot be determined Errorcode 2960) - I have 2 indexes in an inactive state and I think that these messages relate to those indexes.

When I assign a system ID for the Indexmanagement Taskqueue Reader Task as:

KM -> Content Management -> Global Services -> Scheduler Tasks -> Indexmanagement Taskqueue Reader

Under CM systems, I see entries for 2 numbers. I do not know the significance of those 2 numbers, don't think that these refer to installation numbers.

Can any one tell me what is the appropriate value for CM System ID that needs to be entered here?

[KarstenH · Accepted solution]

Hi Phani,

have you opened a message to SAP support?

Regards, Karsten

[Former Member · Accepted solution]

Hello Phani

When I was having this problem, I selected both the CM Systems available in the Index Management Task Queue Reader. As soon as I did that, indexing worked once again for me, and by doing this I was assured that TREX was indexing all CM systems available.

I'm afraid I can't shed light on what those numbers actually refer to specifically, however.

Hope this helps,

Fallon

[former_member195891 · Accepted solution]

Hi Phani, hi Will,

The number shown there the node numbers of the J2EE server nodes available in your landscape. If you have e.g. a single server with 2 nodes you will see the two node IDs. There are several ways to get the IDs, e.g. by using the url http://<server>:<port>; -> System Information.

Please do not assign any task to more than one node, otherwise the same task will be executed on every node.

Kind regards,

Roland

[Former Member · Accepted solution]

Hello All,

I have not yet opened a message with SAP as it was an evaluation box. I am still at the point where the index status is active, queue status is idle & the documents are in the state "To be Synchronized".

Thanks everyone for their feedback.

Regards,

Phani.

[Former Member · Accepted solution]

Hi Phani

If the documents are in the "To be Synchronized" state, flushing the queue usually works. Very odd that it's not... Are you doing that from within the TREX monitor? Click on the 'Change Queue Status' link, select your index, and click 'Flush.' This should push them to the end of the indexing process. Otherwise, I'd also recommend an OSS note.. or try searching current SAP notes for a similar issue?

Hope that helps,

Fallon

P.S. Thanks for the information about the server nodes, Roland!

[Former Member · Accepted solution]

Hello!

You said you see two CM-Systems in your Landscape Definitions. Do you really have two active server-nodes?

I'm asking that because I ran into troubles with indexing after a system copy from one host to another.

During the copy/installation-process a new CM-system is generated but the old one is NOT deleted!

As FollowUp-Activity of the systemcopy procedure you have to delete the old system within System Landscape Definitions after the installation on the target system has finished.

If you do not so, maybe there are some tasks are assigned to the "dead" system which prevents TREX from function probably??

Kind regards,

Gerhard

[Former Member · Accepted solution]

Phani ,

Did you resolved this issue??I am facing the same issue of documents in "to be synchronised " state, flushing doesn't work for me as well.Can you please tell me how did you resolved it??

Thanks

[Former Member]

Hi Phani,

You can check the application log and also under System Adm -> Sys Conf ->Index Adm -> Trex Monitor -> Display Queue Entries -> select ur index and check if it is giving the status as preparation failed or any other error, it will wpecify you the error code and reason.

This might help you.