Part 3: internal MS Exchange plan: tape-free backup
- TAGS:backup tape, cluster, clusters, collaboration, collaboration software, disk, disk array, e-mail, email, enterprise, Exchange, Microsoft, Microsoft Exchange, MSFT, raid, replication, sas, SATA, serial ATA, tapes
- IT TOPICS:Applications, Cloud Computing, Data Center, Enterprise Apps, Hardware, Infrastructure Management, Servers, Storage, Windows
By Richi Jennings. June 17, 2010.
Got Microsoft Exchange? Earlier, I started talking about how, at its Tech-Ed IT-fest, Microsoft last week offered some fascinating insights into its internal use of Exchange. Today, I'll talk about how Microsoft IT has ditched traditional tape-based backup, in favor of cheap disk arrays.
As a recap, here are the highlights of what Microsoft IT said:
- Not using disk arrays, but direct-attached storage
- Not using RAID or clustering, but continuous replication onto spare servers
- Not using tape backup, but cheap disk arrays
If you missed either of the previous two parts of this series, you can click on the two links above. Go ahead, I'll wait...
Today, let's look closer at the third of these radical differences from MSIT's previous server architecture...
No tape backup: as Microsoft's internal message stores grew, backing up to tape became more and more problematic. Tape backup performance wasn't keeping up with the increasing demands on it. It got to the point when there weren't enough hours in the day to perform a full backup of a typical message store. Also, the tape subsystems that were needed to achieve previously-acceptable levels of performance were expensive: both in terms of the tape drive hardware and of the blank media itself.
Something had to give. As we saw last time, MSIT already decided to redesign for high availability without using a shared storage cluster architecture. They designed hot-spare stores in case of disaster, kept updated using Exchange 2010's replication functionality. This in itself fulfilled one of the tasks for which traditional tape backup was used.
However, another requirement for message store backup is to be able to roll back to earlier backups. Having a hot spare isn't sufficient -- what if the database got corrupted? The risk is that the hot spare would also be corrupt.
What MSIT decided to do was backup to a third level of nearline disk. This time, the backup target was an inexpensive RAID5 array, made up of SATA drives. Nearline SATA units are only a litle more expensive than consumer drives, but are typically a little faster and more reliable.
The backup regime called for incremental stores every 15 minutes, plus a daily full backup, taken overnight. MSIT are able to keep at least a week's worth of backups on spinning media, in case they need to roll back to a previous point-in-time.
For performance, the backup is driven from the hot-spares, not from the live databases. This makes sense because it doesn't cause extra I/O load on the live databases, just on the spares, which would otherwise be idle (except for the replication tasks).
Sounds crazy, but I can't think of a better way if your objective is to do more with less, your SLA calls for 99.99% uptime, and your per-user quota is 5GB of email. This is some unconventional thinking that other Exchange shops can learn from, even if they're not as big as Microsoft.
If you missed the previous parts of this series, you can click below to catch up:
Â
How about your mailbox role backup architecture? Leave a comment below...
 |
 | Richi Jennings is an independent analyst/consultant, specializing in blogging, email, and security. A cross-functional IT geek since 1985, you can follow him as @richi on Twitter, pretend to be richij's friend on Facebook, or just use good old email: TLV@richij.com. |
You can also read Richi's full profile and disclosure of his industry affiliations.
Free Insider Guide