Right now it is not clear what causes the long runtime so there is no way to tell if any parameter setting might help.
I don't see how partitioning would help here - after all it is one single hierarchy that gets build up here and there is no way to know upfront, which parts of the hierarchy are stored in which partition.
Did you check CPU/memory utilization during the HierarchyCreate ?
I might miss explained
User SYSTEM has CATALOG READ grants and I can see the sequences under catalog/<my schee>/sequences
I can't see the sequences under catalog/public synonyms/M_SEQUENCES although for exampel i can see the M_TABLES synonyms
so the issue seems related only to sequences.
When I created another user and logged in as SYSTEM and granted the user CATALOG READ it didnt help
I tried pressing "Cancel" in the progress bar in HDB Studio and the process just stayed in a state of waiting to cancel. The following did got work for me either:
ALTERSYSTEM CANCEL SESSION'400210'
However, when I did a disconnect, I at least saw it disappear from the progress bar/tab in HDB Studio and then was able to do other queries.
altersystemdisconnectsession'400210'
It appears that under Performance --> Sessions, this particular session is still waiting to be canceled.
Hi , I'm researching about performance of system replication.
I can't understand about log ship timing from some of docs.
About sync replication mode , It works like following ?
So it means "Sync mode need x2 transaction time + network delay" .... I doubt this .
Another idea , It works like following ?
Which is close current HANA work ? Just I want to know Sync mode need x2 transaction time (from single node) or not .
Regards,
JT
And How about ASYNC ? Does anyone measure one transaction performance on single environment and replication environment?
Hello Friends,
In our scale-out HANA DB landscape, we have a total of 4 nodes and one is being SLAVE node. The landscape is to support the BW Production application.
When I tried to run the "Optimize Table Distribution" option under HANA STUDIO --> Landscape --> Redistribution, the Save Current Table Distribution is show the Status as FAILED with Error: "sql syntax error".
Is there a reason, why the Redistribution run is failing? What exactly is the "sql syntax error" and where can I find more details about this error?
Thank you,
Sri
Another question relating to the case
I've opened a case with CSS
They acknowledged the problem but refuse to give support stating
"...
Please provide the SISM information of the system relating to DLM
service organization.
We are not authorized to work on system if system does not belogs to
DLM.
..."
I got my system from MyLab, any one knows whats the SISM & DLM for those VM-s ?
Dear Expert!
We are running SAP ECC 6 EHP5 and BI EHP2 on physical servers with Windows 2003 and SQL 2005 environment.
We are planning to move on Virtualization servers then have plan to upgrade and migrate SQL 2005 database to Hana database.
I am seeking expert advise what types of hardware should require for the Virtual environment then migrate to Hana database.
Waiting expert advise.
Anwer Waseem
Hello,
For some tables in our HANA the load into memory starts automatically on startup.
How do we configure which tables to load into memory on startup?
Thanks,
Amir
As I've described in Richard's and my book (Book Announcement: SAP HANA Administration) there are basically two ways that column tables get re-loaded during the db restart:
- PREloading, where you have to actively set the PRELOAD flag for the tables (check the ALTER TABLE command reference- there is not GUI option for that) you want to have in memory right after instance startup
and
- REloading. This is a automatic mechanism that has been implemented for replication sites, so that in case of a fail-over the required data is already in memory. For this SAP HANA automatically keeps track on what columns are loaded and reloads those after instance restart.
Cheers,
Lars
Hi Lars,
good that you brought this point. I was searching for PREloading the parameter in configuration tab. But couldn't find it.
1. Was it recently removed?
2. Why was it removed from GUI!
3. And now if don't want my 'any' column tables to be loaded during single site startup, how do I stop?
Happy HANA 
Deepak
With SAP HANA only explicitly set parameters are visible in the configuration files.
The GUI only reflects what's in these files (based on contents of M_INIFILE_CONTENTS).
None of that has changed.
You may want to check the documentation for the parameter.
Your last question doesn't make any sense to me.
Hi Thomas et.al
I'm using 'SAP HANA Database 1.00 SP8 (Revision 80) + HANA SAP HANA studio Version: 1.80.3
1. Is there a hook to run a bunch of SQL-s statements after an import on target HANA without user in the loop ?
2. is HDBDD supported in my configuration ?
3. is HDBDD format different than HDBTABLE meaning I'll need to rewrite my schema from scratch (again) in another format ?
How to integrate SAP hana with Pivotal HD Greenplum (HAWQ) database?
Thanks
Sreenivasa Rao.K
What you say is true, but BW has much more sophisticated technology available in The OLAP Compiler in #BWonHANA - which allows for potentially very complex requirements to be modeled. Question is whether the optimizer works correctly for this scenario.
If it doesn't, it's much better to use a graphical calculation view and then consume it into a VirtualProvider. This allows for flexible multi-dimensional modeling.
If you use ROLLUP then you have constrained your model to work with a specific result set. Hope this clarifies.
Sorry Kang I don't quite understand. I thought your requirement was to have the COUNT DISTINCT at a different aggregation level to the measures in the query. Is that not the case?
Delegated measures I thought would only ensure that the measure is not re-aggregated in Webi (which would cause incorrect results).
What Constant Selection does is to allow the COUNT DISTINCT to be at "Country, Region, Store" level even when you have "Country, Region, Store, Banner" in the dimensions.
I'll have to try this for a real-world scenario some time. Two FEMS combined in the OLAP layer is fine if the result-set is small. If the interim-result sets become large then bad things happen in the OLAP layer!
Same with Tom's Blog: #BWonHANA: OLAP Compiler Demo - I'd be interested to check that the COUNT DISTINCT in BW uses the bw-pop for COUNT DISTINCT that was available in SPS06 for SQL. I noticed this month that Analytic View "counters" do not use it.
As you say, worst case for tricky queries, we can create HANA models. The downside is the HANA models are often query-specific.
Update:
We have raised an Support Incident and it is still "WIP". Meanwhile, I disabled Automatic merge on the Table and was doing a Manual merge from the procedure after the loading is done. It has solved the issue. Found the below trace :
[80414]{1601013}[155/108602087] 2014-11-08 22:00:56.191461 e euclid BwPopAggregateParallel.cpp(01064) : Index syncpoint mismatch;object=SCHEMA_NAME:_SYS_SPLIT_TABLENAME~3en(23017)
[80250]{237874}[95/108602134] 2014-11-08 22:00:58.114087 i TraceContext TraceContext.cpp(00702) : UserName=USERNAME
Still waiting for the message from SAP before checking in production.
Regards,
Krishna Tangudu
Anyone ?
Am I in the wrong section ?
Thanks
We could solve this issue: When copying a task from a template, you need to make sure that in the targetquery, the target attribute filename contains the name of the staging table that you want to write to; in the beginning it is filled with a default value.
Hi ,
We want to update SAP HANA revision from HANA SP08 82 to 83.
1. Could you please help with high level steps. { Note : We are not looking for automatic update but manual one }
2. Also which documents from SAP HANA Platform – SAP Help Portal Page should I refer?
Regards,
Vinay