We are running Adabas 6.2 and Natural 6.3 on a Windows 2003 Server Enterprise Edition (physical
server). We need to deploy a clone of this server on a virtualized platform (Xen) in order to be able
to restore Adabas data if the physical server got unavailable.
We are taking backups twice daily, at 15:00 and at 00:15.
In order to have plogs as fresh as needed, I understand we should perform the command "adaopr db=1 et-sync
feof=plog" on a hourly basis, for example. After that, we should copy the latest closed plog to the clone
In case the clone server were needed, we could restore the last backup and process the hourly plogs.
Any ideas, am I missing anything important?
Thanks in advance, best regards
If it works like the online backup (adabck) new transaction cannot start while waiting for all users to come to ET-status.
Users may experience this as the system is hanging.
You could reduce TT but then the users will not have the ‘normal’ time to end their transactions and can get a TT timeout.
Thanks for the answer.
What about perfoming the command: "adaopr db01 noet_sync feof=plog?
Can I copy the diferrent PLOG extents to the clone server instead of the different PLOG files?
In order to reduce the recovery point without freeze the users.
One thing you need to keep in mind when choosing your log switching strategy, is that the regenerate process is different depending on the use of et-sync for the log switch.
When et_sync is used, the completed protection log is for a full database session (i.e. there are no active transactions pending). This means that each plog file is processed in its own adarec job.
When noet_sync is used the plog files are extents of the same database session and must be regenerated in one adarec job. While a single job initially seems simpler, the handling of the multiple input files needs to be looked at carefully as, by default, for each plog extent the utility pauses and waits for the user to input the name of the next file. See the Using Utilities/Adabas Sequential Files section of the Adabas Administration manual for details.
I believe you can still apply these several extents (from noet_sync and belongs to same session) in several ADAREC jobs as long as you do not shut down nucleus.
We’re doing a similar thing - using plog for replication.
Our adaopr command is: