Hello,
I am using SQL Server 2012 SE and we just upgraded the SAN from SAS to SSD. The migration was smooth but we are noticing
failover time has increased drastically (from less than 30 seconds with SAS to 15 minutes SSD). When failover is performed disks are available immediately to the previously passive node which want to become active and also the cluster comes up instantly but it takes very long time for the sql server service and sqlserver agent service to start. There are in total 10 databases and the biggest log file is close to 3GB. The VLF count per database is as show:
| dbname | vlfcount |
| DB1 | 179 |
| distribution | 4 |
| master | 3 |
| model | 26 |
| msdb | 55 |
| DB2 | 275 |
| DB3 | 307 |
| DB4 | 59 |
| DB5 | 47 |
| DB6 | 29 |
| DB7 | 319 |
| DB8 | 32 |
| DB9 | 331 |
| DB10 | 32 |
| tempdb | 67 |
The biggest database size is 5GB. Considering they are not that bad. I performed failover multiple times and whats interesting is the quickest time was 5mins and the slowest was 16 mins for a failover and it never timed out. This is confusing and I was not able find anything in the cluster log or windows eventvwr or cluster eventvwr or sqlserver error log.
How do I know why is it taking so long? Experts need your valuable inputs.