cancel
Showing results for 
Search instead for 
Did you mean: 

BE 12.5 - Exchange disk to tape issue

Khue
Level 4

I am having an issue with backups pausing for a lengthly time. Specifically my exchange backups seem to pause at a certain spot when writing from the b2d file to tape. Currently I am watching my duplicate process chug along at about 1,300 mb/min. I back up 3 storage groups for exchange. Every day I can watch the duplicate process (Duplicate to Tape) pause at approximately 32 gigs and 77 gigs. What's interesting is storage group 1 is about 30 gigs, and storage group 2 is about 44 gigs. This coincidentally corresponds to where the tapes stop.

The big issue with this is that the exchange duplicate job goes extremely fast without these pauses. For example, the rate between these pauses as mentioned before seems to average about 1300 mb/min. With these pauses the duplicate takes about 6 hours while the backup to disk process takes only 2 hours or so. Any ideas what's the deal with the pauses? Do I just have to deal with it or could this be a symptom of something being wrong with the backup job causing the duplicate process to slow down?

1 ACCEPTED SOLUTION

Accepted Solutions

Colin_Weaver
Moderator
Moderator
Employee Accredited Certified

Assuming GRT is enabled on your jobs then there is a huge difference between how the catalog generation operations work between a disk based backup and a tape based backup

As the GRT disk backups have random access available (tapes are serial) and are not written in MTF format then the catalogs don't need to be as full for the disk operation as we can directly reference the data in the IMG folder.

For the tape backup however we have to generate full catalogs for every mailbox which is a longer process.

When duplicated from a GRT set on disk to tape we have perform  that longer process (which will happen at/near the end of each information store)

The above is a very simple explanation of what might be affecting you and it is possible that somethign else is actually causing the problem - but still useful for you to understand that there is a difference in catalog processing between tape and disk backups (for GRT) that might explain it.

View solution in original post

2 REPLIES 2

Colin_Weaver
Moderator
Moderator
Employee Accredited Certified

Assuming GRT is enabled on your jobs then there is a huge difference between how the catalog generation operations work between a disk based backup and a tape based backup

As the GRT disk backups have random access available (tapes are serial) and are not written in MTF format then the catalogs don't need to be as full for the disk operation as we can directly reference the data in the IMG folder.

For the tape backup however we have to generate full catalogs for every mailbox which is a longer process.

When duplicated from a GRT set on disk to tape we have perform  that longer process (which will happen at/near the end of each information store)

The above is a very simple explanation of what might be affecting you and it is possible that somethign else is actually causing the problem - but still useful for you to understand that there is a difference in catalog processing between tape and disk backups (for GRT) that might explain it.

Khue
Level 4

Why can't I get this kind of technical answer for all my questions on how BE works? 

It is GRT. This does seem like a really reasonable theory, therefore you get the cookie.