http://goo.gl/EmKxy0

Archive

Posts Tagged ‘Rep’

Huge Page Support in Linux – To Increase Database Performance.

January 19th, 2014 No comments

Memory Management internally uses TLB cache to map the Virtual address to physical address.
If the TLB cache is small (TLB Miss) (since page size is small), it will need to refer the Page table. Page Table look ups are costly as compare to TLB cache.
That’s reason the applications ( Like Database) which have heavy memory demand can be configured to Huge TLB Pages so that Page Table access can be reduced  and overall application performance can be increased.
Linux has had support for huge pages since around 2003 where it was mainly used for large shared memory segments in database servers.
ASE Database performance can be increased bt 2-7% by using huge page on Linux Platform. You can check Huge Page Support on Linux :

cat /proc/meminfo | grep Huge
HugePages_Total: XXX
HugePages_Free:  XXX
HugePages_Rsvd:   XXX
Hugepagesize:     2048 kB

Source: http://linuxgazette.net/155/krishnakumar.html

From a memory management perspective, the entire physical memory is divided into “frames” and the virtual memory is divided into “pages”. The memory management unit performs a translation of virtual memory address to physical memory address. The information regarding which virtual memory page maps to which physical frame is kept in a data structure called the “Page Table”. Page table lookups are costly. In order to avoid performance hits due to this lookup, a fast lookup cache called Translation Lookaside Buffer(TLB) is maintained by most architectures. This lookup cache contains the virtual memory address to physical memory address mapping. So any virtual memory address which requires translation to the physical memory address is first compared with the translation lookaside buffer for a valid mapping. When a valid address translation is not present in the TLB, it is called a “TLB miss”. If a TLB miss occurs, the memory management unit will have to refer to the page tables to get the translation. This brings additional performance costs, hence it is important that we try to reduce the TLB misses.

On normal configurations of x86 based machines, the page size is 4K, but the hardware offers support for pages which are larger in size. For example, on x86 32-bit machines (Pentiums and later) there is support for 2Mb and 4Mb pages. Other architectures such as IA64 support multiple page sizes. In the past Linux did not support large pages, but with the advent of HugeTLB feature in the Linux kernel, applications can now benefit from large pages. By using large pages, the TLB misses are reduced. This is because when the page size is large, a single TLB entry can span a larger memory area. Applications which have heavy memory demands such as database applications, HPC applications, etc. can potentially benefit from this.

Source : https://lwn.net/Articles/374424/

Memory Mgmt uses Translation Look Buffer(TLB) Cache to map Virtual to physical address, The amount of memory that can be translated by this cache is referred to as the “TLB reach” and depends on the size of the page and the number of TLB entries.

If the TLB miss time is a large percentage of overall program execution, then the time should be invested to reduce the miss rate and achieve better performance.
Using more than one page size(Huge Page) was identified in the 1990s as one means of reducing the time spent servicing TLB misses by increasing TLB reach.
Broadly speaking, database workloads will gain about 2-7% performance using huge pages whereas scientific workloads can range between 1% and 45%.
Huge pages are not a universal gain, so transparent support for huge pages is limited in mainstream operating systems
it is possible that huge pages will be slower if the workload reference pattern is very sparse and making a small number of references per-huge-page.
Many modern operating systems, including Linux, support huge pages in a more explicit fashion, although this does not necessarily mandate application change. Linux has had support for huge pages since around 2003 where it was mainly used for large shared memory segments in database servers such as Oracle and DB2

SAP Sybase ASE Q&A Bank

April 29th, 2013 1 comment

Wait is over Now !! 

Please download the Complete ebook for SAP Sybase ASE Q&A Bank Version 1.0  as below:

Pic

 

 

 

 

 

 

 

 

 

 

 

 

 

Replication Server Interview Questions

December 20th, 2012 No comments

Guys,
Came across below Replication Server Questions, might help you for yours interview preparation. It also includes few ASE questions.
Also Updated the Interview Question Page.
Thanks,

1. What is the difference between MSA and Warm standby ?

2. How can we perform fail over in Case of Warm Stand by and MSA?

3. What will happen, if the queue has data in case of fail over to warm standby?

4. For standby point of view which one is better warm standby or MSA?

5. How can we sync the warm standby database ? If there is some data in queue, what will you do?

6. What is the config parameter you can use to tune a rep agent?

7. Suppose there are 20,000 trans which are getting fail and you need to skip all these trans using error class. Please let me know high level overview to setup
error class?

8. How the multiple replication server can improve the performance as compare to single replication server ?

9. Suppose the query is running slow , how will you investigate?

10. Suppose tempdb log segment is full, how will you troubleshoot?

11. Suppose for a very big database , backup takes 4 hrs to complete . During the backup, due to heavy activity from user log is full, after 2 hours ? What will you do now?

12. How can we guess, in how much time queue will be drain?

13. Suppose queue is getting full, what can be point of bottleneck and why?

14. Suppose log for a database is getting full, how will you trouble shoot? Why log increase is not a better idea? ( Server version is below 15.7)

15. How will you check the latency and what is the latency ?

16. Suppose you want to sync a table in a warm stand replication , How will you do?

17. What are the various factor which can effect rep agent performance?

18. How can you make sure rep agent is running fine ( means scanning is going well in Primary site)?

19. How can we see the number of commands in a transaction in queue?

20. Suppose you have 5000 locks at primary side , and your one delete statement ( which is effecting 7000 rows)went fine. Now in case of Replicate dataserver ( has same number of locks), you are getting out of lock situation? Why it was successful at primary side , what it did not go well in replicate side, despite of primary and replicate have same number of locks?

21. What are the situations when we require to drain the queues?

22. In unix , suppose one dba modifies some of files including dataserver binary to non sybase user, and you want to fix this situation asap? What you will do?

23. Suppose there is thread down issue in replication for target side, and you fetched that its down due to duplicate row? How can we make sure , is it due to duplicate row or duplicate key?

SAP D&T Academy Video – How to Create/Add a Disk Partition in SAP Sybase Replication Server?

December 20th, 2012 No comments

SAP D&T Academy Video – How to Check the Health of a SAP Sybase Replication Server?

December 20th, 2012 No comments

SAP D&T Academy Video : How to Setup table Replication ?

December 20th, 2012 No comments

Are you ready to build solutions that run better with SAP Sybase Databases?

December 17th, 2012 No comments

Smart database technology can help you build products that sell while providing a lower TCO to your customers. Build technically superior products, with database solutions from SAP Sybase.

Read Full story : https://websmp104.sap-ag.de/~sapidp/011000358700001121852012E/index.htm

Sybase CEO John Chen Left SAP After Successful Integration of Sybase

November 1st, 2012 1 comment

SAP AG (SAP) is losing a top executive from Sybase Inc., the largest acquisition in the German software company’s history, after fully integrating the maker of mobile applications for business management

Some Comments :

Chen, i started my career in 1999 as a Sybase consultant and from that time i am following Sybase and still continuing. You have  made a great company and it is a classic story of  rebirth. You deserve full credit to what Sybase is today. I still remember the stock price from $2 when you took over and reaching a peak of mid $30 and then subsequent sales to SAP. I am a optimimist and still believes “SAP & Sybase marriage is made in heaven” . All the best for you Chen ” You are a great leader , innovator , strategist  and a game changert and your story with Sybase  will be remembered in the history books as one of the classics”.  Good luck for your next venture.

 

Source : http://www.businessweek.com/news/2012-10-30/sap-loses-sybase-s-chen-after-gaining-blueprint-for-deals

http://www.bloomberg.com/article/2012-10-30/aElmqt8q7ntU.html

Categories: ASE, News, SAP Tags: , , , , ,

Indian Railways Running on SAP Sybase ASE

August 23rd, 2012 No comments

Indian Railways: Centralizing Ticketing

 

Discover how Asia’s largest rail network is efficiently managing 17 million daily passengers across 6,000 stations with computerized ticketing. With Sybase database technology, Indian Railways is enhancing customer service with 24/7 ticketing and transaction times of less than 20 seconds per ticket.*

Full Story ::  http://www.sap.com/solutions/technology/database/customer-reviews.epx

Implementation of Function String in Sybase Replication Server(SRS)

January 30th, 2012 No comments

These experience shared by  Senior DBAs as name mentioned, Hope this  will help you to understand more about function string from implementation point of view in a Replication environment: 
Craig Oakley , Senior DBA.
—————————-

We used function strings when we wanted to replicate all columns to some servers, and only selected columns to other (web-facing) servers. This was particularly useful before Rep Server allowed multiple RepDefs on the same table. One concern was text columns which were not being replicated to the web-facing server: we had to create a function string to get a text pointer (we used a one-row table and just update all the text columns on top of each other, as the value was not needed on that server): failure to get a text pointer cause the DSI to go down, and we could not specify that as a condition to ignore.

Beyond this, I would imagine function strings could help specify how you want the update to be done, which could be a performance improvement. It would also allow for a different implementation at the replicate than there is at the primary (such as a table at the primary being two joined tables at the replicate).

 

Sukhesh Nair, Senior Sybase DBA
———————————–

We used to have a setup where data was replicated from sybase to oracle as also to a warm standby sybase server. Rep Server function strings helped in filtering data that would need to be passed to Oracle. It helped immensely in streamlining the data flow to targets by manipulating the incoming data through function string. I feel it is one of the most advanced and useful yet very less used capabilities of Sybase Rep Server.
The deterrent could be because of the complexity it would introduce to the replication system. The setup we had worked wonderfully and never gave us any major problems. Without proper monitoring (which needs to be scripted by DBAs) it used to be hard to maintain. Many of the current Rep Server administrators I see do not have adequate knowledge or experience of handling function strings.

Rey Wang , Senior Sybase DBA
————————-

You can map the delete to no op with functional string.

Partha Gogoi Senior DBA
————————-

We use function strings to transform data at the replicate..We have databases being replicated from Toronto and New York to London, Sydney and Singapore and the client ids are transformed at the replicate because, as per business requirements, the client ids are different at each site.. Of course , having a Universal client id would simplify things , but the systems and databases at each site grew independently until replication was set up and it would be a lot of rework to change all the client ids at the replicate sites

Øystein Grinaker Senior DBA
—————————

A Function String could be used to change default behaviour.
Say you delete a row in a table on PDB, but you do not want to delete the row on the RDB. Then make a change in rs_delete. You may make the rs delete just to make a logical delete by updateing a deletemarker for that spesific row.

 

Source : Linkedin.com

http://www.linkedin.com/groups/What-are-possible-usage-Function-3955841.S.65585220?qid=4eb177c2-d7e3-4572-9be4-92d640046c6c&trk=group_most_recent_rich-0-b-ttl&goback=%2Egmr_3955841