We have installed High Availability between 2 systems in bidirectional mode. We have problems with application part, not between collections.

We have replicated $BSE folder and excluded in them this folders
$BSE/bin/*
$BSE/log/*
$BSE/lib/*
$BSE/tmp/*
$BSE/tools/*

There is any other folder that I have to exclude?

Thank you in advance,

Nobody has High Availability?

"High availability" is not a catch-all phrase I'm afraid. If you post some details about the issues you're having you might have better luck in getting a response. :)

Thank you very much for your suggestion. I explain we want to do.

We have 2 systems with 2 data base. In one of them we have the middle of the users and in the other one the other middle. We want to connect the 2 systems with "High availability".

The "High availability" serve for have the data of the machines synchronized, that to say, if I write a transaction in one of the machine, automatically the data appears in the other machine.

We have no problem with the data collection, but we have problems with the program. We know that if we change a program of if we convert to DD a table we have changes in:
/baanERP/bse/application
/baanERP/dict

but for example if we change a user it don´t work, because there is something that change in /baanERP/.... (I don´t Know) that we have to send to other machine.

Somebody know the parts of /baanERP that never change and the parts that change if you make a change in the program?


Thank you in advance,

From the list of the previous post.
You should include the following directories
${BSE}/lib
${BSE]/tools.

Please also share how you are synchronizing transactions at the Database level, based on that you can further decide which subdirs of ${BSE}/lib should be included for FS replication/synchronization.

We use an application whose name is Mimix.

In AS400 the collections are stored in something called Libraries. In each library there are all the baan tables. The reply/synchonize is made with the changes maked in the data of this object. I think this process is independent of Baan, because after include a register in a table I see the change in the other machine.

All the table that begin for "tt" is stored in another collection, and this part is being synchonized but the result of the change I don´t see, for example, a change in a user.

Hi Monica,
We also using "Mimix" to synchronize the data between two AS/400 servers (570 servers-Production & 840 server - HA sever). Keep in mind never synchronize /$BSE/tmp....you end up with issues in printing. HA is not small task, you need to talk to Baan, Baan customers who are doing and who had a good knowledge about...that is the only way you get through it.

Good luck.
Thanks,
Mur

Hi mraguthu,
We have spoken with Baan, but they haven´t said us which of the foder in /baanERP we have to synchronize.

Can you say us the structure of IFS (WRKDGIFSE) of /baanERP?

Thank you in advace,

We are syncronizing all folders in IFS expect $BSE/tmp and some "tmp" files. Because tmp files will be locked while Mimix process. (that is not good for production system)

Thanks,
Mur

Normally all the changes to these relevant folders are updated only by techncial team . So for all technical people we have the a startup session specified in this thread on replication of FS objects (http://www.baanboard.com/baanboard/showthread.php?t=10551) .

Also this will help u exercise some latitude for FS syncronization like user data , remote user data , tabledef6.1 etc .

As this program has to track only the changes of developers , it's incremental and very minimal load on the OS resources .

We are also using a kind of HA Solution. I try to describe:

The whole system is a cluster of 7 servers and 2 loadbalancers.
2 Servers are HA for the database and some "hardcore" users, 5 (smaller, SunFire V240) servers are the application server. All severs are sharing the $BSE-path by NFS (it is on a SAN). Normally one of the 2 bigger servers is for the DB and the other one works as application server for "heavy" users (warehouse and jobs). We don't have production, so warehousing and order entry is our biggest thing.
The loadbalancer are distributig the "standard" user to the application servers (two because of HA).

I think the most interesting part for you is sharing by NFS (or something equal on AS400).

Maybe it helps

Regards
csecgn

Thank you very much. We finally are synchronizing:

/BAANERP/BSE/APPLICATION/*
/BAANERP/BSE/LIB/USER/*
/BAANERP/DICT/*
/BAANERP/BSE/LIB/DATECURR

There were some files more but we don´t change any of this file after installing baan and I think in our system isn´t necessary.

Apparently it works fine.
Again thank you very much.