Please start any new threads on our new
site at https://forums.sqlteam.com. We've got lots of great SQL Server
experts to answer whatever question you can come up with.
| Author |
Topic |
|
gator8869
Starting Member
37 Posts |
 Posted - 2007-11-19 : 14:54:54
|
| Recently we upgraded the drives on our SAN that holds our clustered sql server data. The migration to the new drive seemed to work fine, but then I started using the old drives to store backups etc.. I added the new drive as a dependency to the SQL server cluster resource and created a new maintenance plan to do a full back up at 12:30 am and 12:30 pm, with trans log backups between, all going to the new drive. Well the 12:30 am full backup failed and at the same time the cluster failed over to our second sql server. I thought it might just be a coincidence. I moved the cluster back to the first sql server. The tran. log back ups and the 12:30 pm full backup ran fine, but again that night the 12:30 am backup failed and SQL failed over to B, this happened two more times since. Why does the backup cause the cluster to fail sometimes, but not others? The tran. log backups have never caused it to failover. Any ideas?Here is some info1) Windows 2003 Enterprise with SQL Server 20002) I didn't perform the drive upgrades, we hired that out, but the level of service we received was pretty poor. He changed the cluster groups around a bit and added the MSDTC cluster group in with the SQL cluster group. He also didn't add the second drive as a dependency to the sql server resource, I had to figure that out on my own.3) Here are some of the errors in the Application event viewer18210 :BackupDiskFile::RequestDurableMedia: failure on backup device 'S:\MSSQL\Backup\TIERONE_db_200711180030.BAK'. Operating system error 21(The device is not ready.).18210 :BackupMedium::ReportIoError: write failure on backup device 'S:\MSSQL\Backup\TIERONE_db_200711180030.BAK'. Operating system error 21(The device is not ready.).4) Here are some of the errors in the System event viewer right after the backup tries and fails.A failure due to grouping might have been encouneterd. A Mutually exclusive fault between paths to the host might prevent the luns from being able move to the correct controller. This problem can be fixed by restoring the paths which are broken.The Driver has detected a path failure to Subsystem ID 500805F3000F6A60. Dump Data 0 contains the Phys Path Info < Port | Bus | Target | Lun >. Dump Data 1 contains the Driver Status. Dump Data 2 contains the HBA Slot Number (ffffffff if unavailable). Dump Data 3 contains extended Driver Status.The device, \Device\Harddisk1, is not ready for access yet.Reservation of cluster disk 'DISK S' has been lost. Please check your system and disk configuration.Cluster resource 'DISK S' in Resource Group 'OKWCSQL' failed.============================I have went through several articles and can't come up with much. I've looked over all the dependencies of the cluster and everything seems to look correct. Any ideas would be appreciated.Thanks,Jeff |
|
|
tkizer
Almighty SQL Goddess
38200 Posts |
Posted - 2007-11-19 : 14:59:18
|
| MSDTC should not be in the SQL cluster group.Does the disk resource show as online in Cluster Administrator?Tara KizerMicrosoft MVP for Windows Server System - SQL Serverhttp://weblogs.sqlteam.com/tarad/ |
 |
|
|
gator8869
Starting Member
37 Posts |
Posted - 2007-11-19 : 15:08:09
|
| Yes, I know best practice is to move it to it's own group, or to have it in the Quorom group, but that is what this "professional" did. I may eventually re-create the group with its own drive etc., but I was hoping to get this issue fixed first. Everything shows as online, and the drive is accessible from SQL, I can run manual backups, or invoke the maintenance plan and it will run without failing, it just hasn't been successful at the midnight run. The drive in question was our old exchange raid, just wiped and re-assigned to SQL. |
|
|
|
rmiao
Master Smack Fu Yak Hacker
7266 Posts |
Posted - 2007-11-19 : 23:04:51
|
| May need apply latest firmware. |
|
|
|
TRACEYSQL
Aged Yak Warrior
594 Posts |
Posted - 2007-11-20 : 18:09:54
|
| Not sure if this helps but can you see this path backups in the cluster ? i know in order for me to use san i have to be able to see it in cluster they tell sql there is some new directory i believe in the resources...part.Otherwise i cannot use it. |
|
|
|
|
|
|
|
|
New drive on Clustered SQL server