DataDomain restore speed for Exchange

[myrdin]

HI,

i am doing some tests after playing around with a DD2200, and i am a bit confused.

I have a Exchange box with about 500GB of edb file in it so very small. It gets backed up pretty quickly on a ddboost rep, i can get to 100/150MB/s.

I am trying to mount the db through Veeam explorer (Restore exchange items), it is running now for more than HALF AN HOUR, and it still running. Now, if i look at DD it is basically doing NOTHING. IOPS are very low, network very low, cpu low everything low, which brings to the conclusion that the DD is basically doing nothing. Why?

Mounting a file server (2TB big), takes about 1-2 minutes which is acceptable.

My first thought that the DD was the bottleneck, but since i cant see nothing is happening (i can see a lot of traffic during backups), is this actually Veeam? thoughts?

[myrdin]

i have actually stopped the process, now i got an error: loaded mount is not found. Thoughts?

[chrisdearden]

I dont suppose you have a lagged copy of a database anywhere ? The log replays can cause delays in the VEX interface opening. Best way to test it would be to mount the Volume for FLR and copy the EDB/Logs off to a staging area and see if VEX mounts straight away or with some delay.

[TheHannu]

Hello,

was there any resolution/clarity on this one ?

Have very similar issue with VEEAM, DD2500+Boost + Exchange restore
Have checked integration and best practice document (Veeam+EMC) and those should be in order...

In my case mount succeeds, but takes like 20+ hours.
When doing Backup directly to disk mount (staging area on primary storage) it takes ~5 minutes to mount.

Thanks for the help!

[foggy]

Please contact technical support. Logs review should tell what operations take longer than expected.

[TheHannu]

Thanks for the quick reply.

Will engage support (VEEAM/EMC), but was just streching my luck to see if this very similar looking case already had some colsure. (or some simple thing to do/not to do)

Cheers!

[veremin]

A lot of issues are environment specific, and it's really hard to find out those without seeing actual logs and infrastructure. That's why forum correspondence is not always a panacea.

Once ticket is opened, kindly, post its number here for the convenience of future readers.

Thanks.