Welcome to the new Gigaspaces XAP forum. To recover your account, please follow these instructions.

Ask Your Question
0

Instructions for migration from Gigaspaces 6.0.3 to 6.5

We are researching migration from Gigaspaces 6.0.3 to 6.5, however the page at http://www.gigaspaces.com/wiki/displa... does not have any instructions; can you provide this?

{quote}This thread was imported from the previous forum. For your reference, the original is [available here|http://forum.openspaces.org/thread.jspa?threadID=2536]{quote}

asked 2008-08-18 11:18:26 -0500

wasefmasood gravatar image

updated 2013-08-08 09:52:00 -0500

jaissefsfex gravatar image
edit retag flag offensive close merge delete

4 Answers

Sort by ยป oldest newest most voted
0

See here changes with 6.5 API: http://www.gigaspaces.com/wiki/displa...

The main differences: - Listening Port - it is now 4062 - Space , space container and cluster Schema files exists only within the JSpaces.jar. - Perst DB is not supported

In most cases there is no need to change your code. To be on the safe side compile it with the new 6.5 jars.

Shay

answered 2008-08-19 05:14:48 -0500

shay hassidim gravatar image
edit flag offensive delete link more
0

Thanks. Are there any (for 6.5): 1. Benchmarks available 2. Examples of wildcard search (using SQLQuery etc).

answered 2008-08-21 12:21:18 -0500

wasefmasood gravatar image
edit flag offensive delete link more

Comments

  1. Yes - there are benchmarks for 6.5. Send me email to shay at gigaspaces.com .
  2. Yes - see: http://www.gigaspaces.com/wiki/displa...http://www.gigaspaces.com/wiki/pages/...

Shay

shay hassidim gravatar imageshay hassidim ( 2008-08-22 07:05:56 -0500 )edit
0

Can you provide some explanation on how GS 6.5 process works for wildcard queries when GS is connected to a backend database? Do wildcard queries behave any differently when GS is connected to a backend database as compared to without a backend? Does GS use the backend's SQL engine to query the data and load the results, and merge it with the results from the data that is in the in-memory cache (and not yet in the database)?

SM.

answered 2008-08-25 09:05:54 -0500

wasefmasood gravatar image
edit flag offensive delete link more

Comments

When running in LRU mode having EDS the space queries the space first. If the amount of matching objects within the space are less than the amount of max_objects specified as part of the readMultiple parameter (phase 1) the EDS is queried (phase 2). This result is loaded into the space and used as the result set send to the client.

The above is relevant for all queries conducted with the space.

Specifically for wildcard queries the matching process does not use indexes , so the objects within the space are scanned. This means that as much as you will have more objects within the space it will take more time to complete the first phase of finding matching objects within the space. That is the reason why we recommend when having very large amount of objects within the space and much more data stored within the database to query the database directly bypassing the space wildcard query mechanism. This should use remoting (or the executors with 6.6.) that will delegate the query to the database in transparent manner and return the result back into the client that will have the appropriate reducer.

Shay

shay hassidim gravatar imageshay hassidim ( 2008-08-25 10:09:11 -0500 )edit

I am using 6.5 since that is the latest production release. Can you point me to some examples of the "remoting" mechanism that will delegate the query to the database in transparent manner and return the result back into the client

Can you provide some statistics on the numbers at which it would be good to directly query the database. "That is the reason why we recommend when having very large amount of objects within the space and much more data stored within the database to query the database directly bypassing the space wildcard query mechanism"

Edited by: s m on Aug 27, 2008 2:07 PM

wasefmasood gravatar imagewasefmasood ( 2008-08-26 11:49:24 -0500 )edit

I would say that once you have more than 200K-300K objects within the space I would delegate RegEx query to be done with the database directly. Scanning such amount of objects would take about 3-5 sec. I guess this is where it would make sense to delegate it to the database.

Simple remoting example located here: http://www.gigaspaces.com/wiki/displa...

You can extend this very easily to have a generic Query command object sent to the space to delegate special type of queries to the database and others to be performed with the space.

Note that starting with 6.6 the right way to implement remoting should be done via Executors: http://wiki.gigaspaces.com/display/GS...

Shay

shay hassidim gravatar imageshay hassidim ( 2008-08-28 11:10:26 -0500 )edit

Your Answer

Please start posting anonymously - your entry will be published after you log in or create a new account.

Add Answer

Question Tools

1 follower

Stats

Asked: 2008-08-18 11:18:26 -0500

Seen: 76 times

Last updated: Aug 25 '08