With Backup Exec 2010 or PureDisk you can have the deduplication happen on the media server. Say this happens, now you have the deduplicated backup set. You now want to replicate this backup to DR. With PureDisk I believe it will now only send the changed blocks over the wire that are not already at the DR site’s PureDisk Storage Pool. Is my assumption about how PureDisk works correct?
Will this work the same way with Backup Exec 2010? Or will it not compare the blocks that are already existing at the DR site and will just send the whole (locally deduplicated) backup to the DR site?
he Administration Guide is not clear. It just seems to me that it is a full copy. Page 1536 goes into detail. I have added the small subset of the
For your first question, I'm going to defer to Symantec and what they recommend ;)
For your next, My guess is that you'll be more CPU bound than disk bound in general.
My take is a 64bit server, quadcore dual socket CPU, 8GB of RAM minimum... Then for disk, large sector/block sizes, disk aligned, write back cache, and RAID5 or 50... Not a fan of 6 for the write penalty, and 10 is for Db's not backup data IMO.
In Backup Exec, you will be able to run a duplicate backup from Puredisk on media server to puredisk on another server, and it is optimized to transfer deduplicated data. So, it is like global deduplication.
No, you can have puredisk implemented on the media server and you may have another puredisk server at a different location. So, with dedup option on Backup Exec, you can implement PDDE on media server, which will be a dedup storage folder and run a duplicate backup job to another PD server as an OST device.
You can go media server to PureDisk node, media server to media server, or Open Storage device to Open Storage device.
The data that is transferred is ONLY the unique blocks of data that has changed.
Remember, PureDisk is a lot more scalable supporting up to 32TB of data (per node up to 16 nodes), and BE only half that (with no clustering options).. So for long term retention, a remote Puredisk pool could be beneficial... And the PureDisk starter bundle makes pricing just about inline to what BE and the dedupe option may cost.
Regarding your statement, The data that is transferred is ONLY the unique blocks of data that has changed, I do not need to implement NetBackup or PureDisk for this small customer. I can have a BE Media Server (deduplication storage folder) to BE Media Server (deduplication storage folder) and It will only transfer the unique blocks that are already not at the DR site, right?
But then this brings up a few more questions :)
Server-side deduplication (with Granular Recovery Technology enabled) is supported for Exchange and SharePoint. However Optimized Deduplication is not available for backup sets that are enabled for Granular Recovery Technology. Without Granular Recovery on SharePoint is there any way to restore individual documents from the backups?
You can only create one deduplication storage folder on a media server (Ref: Admin Guide pg 1524). Does this mean that if you want to create a 16TB storage pool, you need to create a 16TB LUN on the Windows Media Server? Or would you have multiple LUNs stripped and create a Dynamic Disk? Also, if you want to add more storage, will Backup Exec handle a disk extend (starting with say 10TB and wanting to add 2TB).
In order to copy data from a deduplication storage folder on a managed backup exec server to a deduplication storage folder on another managed media server you need the Central Admin Server Option. If I am using CASO for the sole purpose of enabling this technology, can I just run CASO on the Backup Exec Media Server or should it be on a dedicated node? Also, if the BE server in the DR site is only used for Disaster Recovery Purposes, do I need to license this Media Server?
#1, I want to know myself, this answer.
#2, you can point to a NAS that can grow, or a single NTFS volume. The main thing is that you have a target you can grow to that size, and Symantec is not going to help much here, as you are limited to the OS functions and hardware you have available. My recommendation would be a NAS that has a unified namespace that you can change the backend disks to whatever you need... The Symantec FileStore is interesting and could facilitate this... Cheaper than most traditional NAS devices.
#3, it can be on a BE media server or dedicated node, your preference.. Yes you need another license for the 2nd site as it's up and running and not a cold site.
#4, ?
If I am running a local copy of SQL on the Media Server, should I share the BE Database Storage location with the DeDuplication Storage Location or should I seperate these on seperate physical drives?
Also with 2TB drives out. Do you recommend RAID 10 or RAID 6 for the backend DeDuplication Storage Pool?
For your first question, I'm going to defer to Symantec and what they recommend ;)
For your next, My guess is that you'll be more CPU bound than disk bound in general.
My take is a 64bit server, quadcore dual socket CPU, 8GB of RAM minimum... Then for disk, large sector/block sizes, disk aligned, write back cache, and RAID5 or 50... Not a fan of 6 for the write penalty, and 10 is for Db's not backup data IMO.
As discussed in the thread above, I had mentioned that you cannot use GRT backups with Optimized DeDuplication (deduplication with replication between media servers). The main benefit of the Backup Exec Hyper-V agent is that you can do this single pass backup and then use GRT to restore individual files/folders. Since I cannot enable this with optimized deduplication, this agent is losing most of its functionality. Can you back up the Hyper-V Phyiscal Host without the Backup Exec Agent for Hyper-V?
