Reorg Error appears many times

[Former Member]

HI!

As we accelerated a new cube, I receive Reorg error as soon as I finish building the Index. Also, after attribute change run (daily) in process chain it appears again. We do not have any automation for Reorg.

I know this appears when the indexes are not split equally over the blades.

Wondering if this is a usual case and we should keep fixing it maually using RSDDBIAMON?

Is there any automation setting to do the Reorg?

Please Advise.

Thanks,

Ghassan Matar

[Former Member · Accepted solution]

Jens,

I found that the reorg happens after a process chain that loads data for accelerated cube. At the end of the process chain, it rolls up BWA index and updates explorer indexes. I think I will try to add the ABAP step to Reorganize the index at the end of that process chain.

When I used the transaction RSDDBIAMON and did the reorg, I found in sm37 that it used ABAP program

Step 001 started (program RSDDTZA_REORG_STATUS

or I should use RSDDTREX_REORGANIZE_LANDSCAPE

and what is the difference between these two programs?

Edited by: GHASSAN MATAR on Mar 6, 2011 3:38 PM

Edited by: GHASSAN MATAR on Mar 6, 2011 3:42 PM

[jgleichmann · Accepted solution]

Hi,

RSDDTZA_REORG_STATUS just checks the availability of the BWA.

RSDDTREX_REORGANIZE_LANDSCAPE executes the reorg. You get the alert every time when the alert check is running. This is by default at 0:00. You can change this in transaction TREXADMIN with tab alert -> alertserver configuration or with the standalone tool...

Regards,

Jens

[Former Member · Accepted solution]

Hello Jens,

Do we need to schedule this report on a Daily basis or weekly basis?

Thanks for you advice.

[jgleichmann · Accepted solution]

Hi,

which report? the reorg? We are running it once a day. The recommendation is to schedule it once a week it depends on the changes of the indexes. There is a alert check which checks the need of a reorg. Just have an eye on it and schedule it on depency of this alert/warning.

Regards,

Jens

[Former Member · Accepted solution]

Our index size is displayed as 1085MB and in the 'Hin Text" is reorg no.

Do I still need to schedule a reorg job or wait for the Alert monitor tuns RED.

Thanks

SM

Edited by: SM on Apr 19, 2011 3:44 PM

[jgleichmann · Accepted solution]

Currently there is no need to run a reorg, but it´s recommended to schedule it once a week. The reorg isn´t like a table reorg which takes some time. this justs split the index parts over your BWA hosts. It will take in your case abut 1min or less. Just schedule it in a time where no write access is on all indexes, because the reorg locks the indexes for a short time...

You will also free some memory because the reorg deletes also some temporary indexes.

Regards,

Jens

[Former Member · Accepted solution]

Excellent, thanks for your advice.

-SM

[Former Member]

Thanks Jens. That helped a lot.

[Former Member]

Jens,

Yes, you are right. Yesterday, I accelerated a new Cube and I was expecting to reorganize the landscape and it was in my implementation plan. I went to RSDDBIAMON and checked "Reorganize BIA Landscape" and it is gone.

Check Name----

---

Check Description -

---

Action

"Reorg" -

---

Returns yes if reorg is recommended -

---

Reorganize TREX Landscape

Thanks,

Ghassan Matar

Edited by: GHASSAN MATAR on Mar 6, 2011 3:40 PM

[Former Member]

Thanks for the quick reply,

I am just wondering about the consequences of getting this error. If I will do the Reorg Step every day at 9:00 PM and I started to get the error at 10:00 AM. If the error will stay around 11:00 Hours, will it cause any down time or crash or anything that will make the query unaccessible for the end user?

Is there like a time period for this error that can exist and after this period it will cause any damage?

Edited by: GHASSAN MATAR on Mar 4, 2011 2:16 PM

[jgleichmann]

Can you tell us the exact error message? Normally the messages which pop up after an creation are more like 'you should reorg or reindex the named index if you do not so it will result in a bad performance'. So this alerts are not critical and won´t cause any damage. But tell us the error message and we can say how urgen t (why I can´t post the word u-r-g-e-n-t, this checks are really silly, it should be checked for the topic creator but not for others) it is.

Regards,

Jens

[jgleichmann]

Hi Ghassan,

you can use the report which Tansu mentioned. You can create a new step in your process chain to trigger the reorg. Please note that no BWA index should be created at this time. because for a short while the indexes are in read-only mode -> this can abort your process chains. So it is advisable to plan it on a fix time when no (write-) load is on bwa.

There is no rule from SAP which tell you the frequency of a reorg, but once a week is recommended. You can also run it daily, there is no problem if you run it more often...

Regards,

Jens

[Former Member]

Hello

You can use schedule the program called RSDDTREX_REORGANIZE_LANDSCAPE (while there is no BWA indexing is scheduled at the same time ).

Cheers

Tansu