Error !!! ERROR - Database restore utility FAILED !!! (8564)

[qadvietnam]

Hello all,

I try to probkup offline backup Progress DB from Linux Server (OE v10.1B03)
Then I copy the backup file to Windows Server (OE ver11.2.1)

And then I try to prorest db to the instant DB on Windows Server

Error msg, pls help me

OpenEdge Release 11.2.1 as of Mon Apr 29 19:26:32 EDT 2013
Unequal number of user areas detected in the target database. (6770)
Source database contains 37 and target database contains 8. (6771)
Restore failed. (1618)
!!! ERROR - Database restore utility FAILED !!! (8564)

My backup database (mfgproduct.bak) is 4.5GB

Here my log

proenv>prorest c:\qadapps\db\mfgprod c:\backupproduct\mfgproduct.bak -vp -vf
OpenEdge Release 11.2.1 as of Mon Apr 29 19:26:32 EDT 2013
This is a full backup of /u01/mfgproduct/mfgsvr/dataproduct/product/mfgproduct.db. (6759)
This backup was taken Thu Feb 13 13:43:19 2014. (6760)
The blocksize is 8192. (6994)
Partial verification successfully read backup volume. (6765)
Verify pass started. (3751)
CRC check error reading backup block 1 (1147)
CRC check error reading backup block 2 (1147)
CRC check error reading backup block 3 (1147)
CRC check error reading backup block 4 (1147)
CRC check error reading backup block 5 (1147)
Unable to recover previous block in error (1104)
CRC check error reading backup block 6 (1147)
CRC check error reading backup block 7 (1147)
CRC check error reading backup block 8 (1147)
CRC check error reading backup block 9 (1147)
CRC check error reading backup block 10 (1147)
10 read errors have occurred. (6780)

Do you want to continue? [y/n]: n


proenv>
proenv>prorest c:\qadapps\db\mfgprod c:\backupproduct\mfgproduct.bak
OpenEdge Release 11.2.1 as of Mon Apr 29 19:26:32 EDT 2013
Unequal number of user areas detected in the target database. (6770)
Source database contains 37 and target database contains 8. (6771)
Restore failed. (1618)
!!! ERROR - Database restore utility FAILED !!! (8564)

proenv>

 

[Rob Fitzpatrick]

Backup and restore across platforms (e.g. Linux to Windows) isn't supported. Backup and restore across major versions, in this way, is not supported. You're trying to do both.

If the backup was from 10.1B or later (you didn't specify) and if it was from a Windows DB server then it would be possible to use the 10.2B05 version of _dbutil that comes with 11.2.1 to restore the database on your Windows machine, but it would still be a version 10 database and unusable with 11.2.1. You could then convert it to v11 with proutil -C conv1011. But that is not the recommended major-version upgrade path; you should dump and load if possible.

Your only option is to dump and load.

 

[qadvietnam]

My DB is using 7 years until now. The backup file is 4.5GB size.
I try to dump and load before but it takes too much time !!!

Any great solutions from anyone?
Thanks
QAD Vietnam

 

[Cringer]

First solution would be to upgrade the Linux server to 10.2, and the latest service pack. You should be able to do that without having to recompile so it's a no-brainer. Progress have provided a lot of peformance enhancements to dump&load over the years and you're not benefiting from them. Also, 4.5GB isn't that big. Our main DB is ~300GB.

 

[qadvietnam]

First solution would be to upgrade the Linux server to 10.2, and the latest service pack. You should be able to do that without having to recompile so it's a no-brainer. Progress have provided a lot of peformance enhancements to dump&load over the years and you're not benefiting from them. Also, 4.5GB isn't that big. Our main DB is ~300GB.

We can not upgrade OE version to 10.2xx because my customer policy.
Could you pls explain the Bold sentence for me clear?
I try to dump/load before but wait for more 3 days but still no done now, especially tr_hist and gl_hist table ....hic...hic

I have just found this link: http://knowledgebase.progress.com/a...files+larger+than+2GB?&l=en_US&fs=Search&pn=1
is it can help us solve this issue?
Thanks

 

[Cringer]

All I'm saying is that Progress have made the dump and load process a lot quicker in more recent versions of Progress. If you're running 10.1 you won't be getting these benfits.
3 days for a dump of a 4.5GB database means something is wrong, even in 10.1. It shouldn't take that long.
So your customer policy means it's ok to be running software that's ancient, obsolete and basically unsupported? That's not good.

 

[Rob Fitzpatrick]

Let's tackle one thing at a time.

Migration
To go from Linux to Windows you must dump and load, period. You cannot get there with backup and restore.

Dump and load
Three days is a phenomenally long time for a dump and load of a 4.5 GB database. How did you dump the data - Data Dictionary or proutil binary dump?

License
Are you on Enterprise RDBMS or Workgroup?

 

[qadvietnam]

Hi Cringer,
I mean dump DB takes 2 hours only.
Load DB takes 3 days but still finished yet

 

[Rob Fitzpatrick]

How did you dump the data?

 

[qadvietnam]

Let's tackle one thing at a time.

Migration
To go from Linux to Windows you must dump and load, period. You cannot get there with backup and restore.

Dump and load
Three days is a phenomenally long time for a dump and load of a 4.5 GB database. How did you dump the data - Data Dictionary or proutil binary dump?

License
Are you on Enterprise RDBMS or Workgroup?

Hi,
I use Data Dictronary for DUMP and LOAD
OpenEdge Enterprise RDBMS licence
Any ideas from you, pls help us

Thanks,
QAD Vietnam

 

[Rob Fitzpatrick]

Could you pls explain the Bold sentence for me clear?
I try to dump/load before but wait for more 3 days but still no done now, especially tr_hist and gl_hist table ....hic...hic

OpenEdge 10.2B06 and 11.1 introduced substantial improvements to the index rebuild process. This means binary dump and load can run much faster end-to-end than in the past.

See these presentations, and the OpenEdge documentation, for more detail:
PUG Challenge Americas 2012
Still More Database Improvements
Rich Banville
http://pugchallenge.org/2012PPT/NEPUG_Performance.pptx

PUG Challenge Americas 2013
Index Rebuild Performance
Wei Qiu
slides: http://www.pugchallenge.org/downloads/261_Index_Rebuild_Perf.pptx
audio 1: http://www.pugchallenge.org/downloads/audio/261_Index_Rebuild_Perf.mp3
audio 2: http://www.pugchallenge.org/downloads/audio/261_Index_Rebuild_Perf_wav.mp3

 

[Rob Fitzpatrick]

Hi,
I use Data Dictronary for DUMP and LOAD
OpenEdge Enterprise RDBMS licence
Any ideas from you, pls help us

Data Dictionary is the slowest possible method for dump and load. My advice to you is to read the 11.2 Database Administration manual, particularly the section on dumping and loading your data. You should test binary dump and load with a separate index rebuild (i.e. do not automatically rebuild indexes during binary load). You should see a substantial improvement in performance.

You may also want to review this presentation for a good overview of dump and load techniques and optimization:
PUG Challenge Americas 2012
27 Days to 27 Hours: Best Practices in Dump and Load Optimization
Paul Koufalis
slides: http://pugchallenge.org/2012PPT/2012USPUGCh27Daysto27Hours.pdf

Even using Data Dictionary, 3 days is very very slow for less than 5 GB. I am wondering if there are other factors slowing you down, like the quality of storage on your Windows box, other I/O workload on it, the RAID level, a resource-stealing virus scanner, etc. You or your system administrator should take a long look at that box to see what condition it is in and whether it is suitable for database storage.

 

[qadvietnam]

Rob Fitzpatrick,

How about the way to Dump/Load by Data Dict

Data Dictionary is the slowest possible method for dump and load. My advice to you is to read the 11.2 Database Administration manual, particularly the section on dumping and loading your data. You should test binary dump and load with a separate index rebuild (i.e. do not automatically rebuild indexes during binary load). You should see a substantial improvement in performance.

Pls noted we are using OE10.1B03, not 11.2 as you mentioned
As you said, Data Dictionary is the slowest possible method for dump and load. Is there the ways we can dump/load more quick than?
Pls advice us.
Thanks,

 

[Rob Fitzpatrick]

Pls noted we are using OE10.1B03, not 11.2 as you mentioned

Your first post said you tried to take a backup file from 10.1B on Linux and restore it on a Windows box with 11.2.1. Are you not going to load with 11.2.1 on Windows? If you aren't then why did you attempt that restore in the first place?

Perhaps you need to back up a step and clarify exactly what you are trying to accomplish from a business perspective.

As you said, Data Dictionary is the slowest possible method for dump and load. Is there the ways we can dump/load more quick than?

Please re-read my post above, starting with "My advice to you is...".

 

[TomBascom]

You might also want to consider getting some training.