0% Compression with multiple filegroups
epetro
Posts: 69
I have reported an issue with support where, when restoring a backup containing 9 filegroups, only 1 is restored as COMPRESSED.
I have seen this work. In fact, with a backup file from the same production database worked correctly about a month ago.
Since the issue arose, I have upgraded SQL Storage Compress.
The GUI reports a version of 6.0.0.320
The log file reports HyperBac service version 5.5.1.10
The log file reports HyperBac filter version 6.0.0.15
The upgrade and required server reboot did not correct the behavior.
Here is my restore command I run against a Redgate .sqb file.
I tested a scenario where I adjust the move clause for 'EBC_Index' to native file *.ndf. The result was no compression anywhere.
I believe additional logging needs to be enabled.
Anyone know how to bump the logging level (including service/server restarts) and how much overhead is expected for 2 compressed databases?
I have seen this work. In fact, with a backup file from the same production database worked correctly about a month ago.
Since the issue arose, I have upgraded SQL Storage Compress.
The GUI reports a version of 6.0.0.320
The log file reports HyperBac service version 5.5.1.10
The log file reports HyperBac filter version 6.0.0.15
The upgrade and required server reboot did not correct the behavior.
Here is my restore command I run against a Redgate .sqb file.
RESTORE DATABASE [MYDB] FROM DISK = N'K:\DatabaseBAK\MYDB\FULL_(local)_MYDB_noUse_20120209_084922.sqb', DISK = N'K:\DatabaseBAK\MYDB\FULL_(local)_MYDB_noUse_20120209_084922.sqb', DISK = N'K:\DatabaseBAK\MYDB\FULL_(local)_MYDB_noUse_20120209_084922.sqb', DISK = N'K:\DatabaseBAK\MYDB\FULL_(local)_MYDB_noUse_20120209_084922.sqb', DISK = N'K:\DatabaseBAK\MYDB\FULL_(local)_MYDB_noUse_20120209_084922.sqb', DISK = N'K:\DatabaseBAK\MYDB\FULL_(local)_MYDB_noUse_20120209_084922.sqb', DISK = N'K:\DatabaseBAK\MYDB\FULL_(local)_MYDB_noUse_20120209_084922.sqb' WITH MOVE N'EBC_Data' TO N'G:\Database\MYDB.mdfx', MOVE N'EBC_Index' TO N'G:\Database\MYDB.ndfx', MOVE N'EBC_ChargeActive' TO N'G:\Database\MYDB_ChargeActive.ndfx', MOVE N'EBC_Transactions' TO N'G:\Database\MYDB_Transactions.ndfx', MOVE N'EBC_TapeCharge' TO N'G:\Database\MYDB_TapeCharge.ndfx', MOVE N'EBC_LogTables' TO N'G:\Database\MYDB_LogTables.ndfx', MOVE N'EBC_TapeStaging' TO N'G:\Database\MYDB_TapeStaging.ndfx', MOVE N'EBC_EBCTables' TO N'G:\Database\MYDB_EBCTables.ndfx', MOVE N'EBC_Log' TO N'F:\LogFiles\MYDB.ldf', STATS=1, RECOVERYIn this scenario only the filegroup 'EBC_Index' receives the compression.
I tested a scenario where I adjust the move clause for 'EBC_Index' to native file *.ndf. The result was no compression anywhere.
I believe additional logging needs to be enabled.
Anyone know how to bump the logging level (including service/server restarts) and how much overhead is expected for 2 compressed databases?
Comments
After receiving only 1 filegroup with compression, I switched the drive letters (was placing MDF seperate from secondary filegroups) and the inverse received compression !!
I'm not familiar with disk configurations, so I am reaching out to some co-workers.
What I gather:
HP MSA 2000 (shows up in properties as HP MSA2312fc Multi-Path Disk Device)
2 fiber ports
Partition style: GUID Partition Table(GPT)
Write caching enabled