One of the companies I work for(huge, huge Fortune 500) is constantly facing capacity problems and when I am saying constantly, I mean it! It really is almost every week! The whole environment (not that many servers – around 50 or so) uses local SQL Server backups with retention of “n” days. Those backups are stored on both drives where we store the data and the log files(haven’t I told you they do believe that disk space is “cheap”?). I think you can already guess where this is going! The drives that are “holding” our databases are constantly running out of space (getting below the 10% threshold of the total drive space, not that we are facing service interruption).
Here comes the “creativity” part of my blog post. The one that I am not that proud of that much! It was clear to me (after the first 5 – 10 capacity problems) that I need to communicate that with the customer and I need to do it in very effective way! I strongly believe I did it (my other customers are listening when I told them that they not only can, but will face service interruption if we do not extend the drives that is holding their data/log files). However instead of ordering the new drives for example(except when we were about to face service interruption after 5…4…3…2… you know what I mean!), my customer was and continues to actively refuse to do anything and is making me came up with a solutions of the problem. Cool, huh? How easy for him! So these are all the creativity steps I took(and am continuing to do actually) when we hit the 10% threshold:
- Enabling backup compression(if possible) or compressing the files/backup folder on NTFS level(and facing problems if your backup file is more than 30Gigs in size and running on Server 2003)
- Splitting the backups across the disks
- Splitting a specific database backup to more than 1 disk
- Doing the clean-up of the old backups first and then issuing the new backups
- When this was not enough I was going with lowering down the retention period (without telling them sometimes! I confess…)
- Lowering down the retention period again
- Lowering it down again until we hit the current situation (yes, Risk Acceptance was signed for this!) – clean up task fires up and cleans all backups from the day before -> new backups after that – thus we are having just one full + the transaction log backups after the full on the disk at any time. Now, that is a backup strategy! I know you’d agree!
So, that’s what happens when I turn on my “creativity” plan! Please, please, let me know if you have worked for customer like that and you made them realise that they really have to invest in their infrastructure. I want to talk with you!
Thanks, Rick for hosting number 46! I expect to read some really interesting blog posts from the community! Let the Tuesday be…now!