문제

Here's our config:

  1. We have all our data in PRIMARY file group

  2. We have our Full Text Index in another SECONDARY file group which resides on a very fast (but very unreliable) SSD drive. By "unreliable" I mean - it's a Amazon cloud server temporary drive (can be entirely lost in case of emergency restart/hardware failure).

Full-text index is not critical and we're ready to lose it completely as we can obviously rebuild it from scratch.

The question is: is there a way to revive the database in case the secondary file group is 100% lost, without going through the full backup/restore cycle? Like, simply delete the lost file group, then recreate it, then rebuild the index.

(We do have regular full database backups, log backups etc. etc. etc. so we can "restore to the point in time" but this implies long outage)

P.S. Also, any thoughts on how viable our approach is would be very much appreciated.

도움이 되었습니까?

해결책

(answering my own question)

OK, so we ran a simulation - what will happen if we lose the secondary filegroup file?

Not good.

The database gets stuck in recovery mode and you can't do anything about it - you can't set the "emergency" mode, you can't change it to "single_user" mode - nothing. Basically, every command you run throws an error Unable to open the physical file "xxx.mdf". Operating system error 2: "2 (The system cannot find the file specified.)

The only solution is to restore the database. But "rebuilding" a secondary filegroup won't work, even if there's only a fulltext index in it.

라이센스 : CC-BY-SA ~와 함께 속성
제휴하지 않습니다 dba.stackexchange
scroll top