Redesign v5 Replication to v6 using silver-peak

[JoeIanni]

Hi,

I am currently replicating across 2 Mbit Wireless WAN, with ping times varying from 25ms to 100ms. Maximum through put I can get is approximately 1.7 - 1.8 Mb. We have a couple of Silver peaks at each site and are replicating ESXi to ESXi. We are getting around 90% - 93% reduction on a daily basis with . The Veeam v5 server is at the source site (Which is the not the optimal for version 5) running in Direct SAN mode using iSCSI along with DAS for backups as well.

After reading the user guide , I understand it would be beneficial to install a backup proxy at the Target end? I have a few questions regarding the redesign of this.

Source Site,

1. When I upgrade to v6 should I need to split up the current backup server and run it as a proxy server only or can I keep it as a proxy/repository/backup srver? Take into consideration that we are replicating 15 VM's and backing up over 35 VM's. There was no mention of what is considered small / medium / large

2. Is there any recommended settings for WAN accelerators other than removing compression from Veeam which is currently off?

3. Has there been any update to backing up and replicating at the same time?

Target Site,

1. At the target site is it better to have a VM as the backup proxy or a physical?

2. How much space should I cater for with respect to snapshot size?

Regards,
Joe

[foggy]

1. When I upgrade to v6 should I need to split up the current backup server and run it as a proxy server only or can I keep it as a proxy/repository/backup srver? Take into consideration that we are replicating 15 VM's and backing up over 35 VM's. There was no mention of what is considered small / medium / large

1. At the target site is it better to have a VM as the backup proxy or a physical?

Please review the sticky FAQ topic for considerations on physical/virtual proxies/backup repositories.

2. Is there any recommended settings for WAN accelerators other than removing compression from Veeam which is currently off?

Also please search our forum for existing discussions on using different WAN accelerators: http://forums.veeam.com/search.php?st=p ... ccelerator

3. Has there been any update to backing up and replicating at the same time?

v6 allows parallel backup and replication in terms of processing the same VM by different jobs at the same time.

2. How much space should I cater for with respect to snapshot size?

The space required to store restore points really depends on the amount of changes your VMs provide between replication cycles.

[tsightler]

Source Site:

1. When I upgrade to v6 should I need to split up the current backup server and run it as a proxy server only or can I keep it as a proxy/repository/backup srver? Take into consideration that we are replicating 15 VM's and backing up over 35 VM's. There was no mention of what is considered small / medium / large

It's really difficult to answer this question since it's not all about the number of VM's, but also size and change rate and how often you backup/replicate. I would say that, assuming you are doing each of these once a day, I'd just keep one server.

2. Is there any recommended settings for WAN accelerators other than removing compression from Veeam which is currently off?

Actually, if your WAN accelerator solution includes "caching" you might try with compression. My initial testing with compression enabled in V6 would indicated that our default compression, using WAN target, will typically provide a 50-80% reduction on it's own, while the WAN accelerator will still provide an additional 50-70%. Overall this will still only put you in the same 90-95% range you are already seeing, but will pollute the cache less, leaving more room for non-replication traffic in the cache. I reserve the right to change this opinion after I complete more testing.

3. Has there been any update to backing up and replicating at the same time?

With V6, if two jobs attempt to process the same VM at the same time, the second job will skip that VM and continue with other VMs within the job. It will then retry the skipped VM's until they are available to be processed.

Target Site,

1. At the target site is it better to have a VM as the backup proxy or a physical?

Honestly a VM makes an excellent target proxy since it can leverage hotadd, however, at your link speed, it's not really going to matter much. Actually, network mode may very well be faster because there's a lot less setup overhead that hotadd.

2. How much space should I cater for with respect to snapshot size?

Impossible to answer since you didn't say how many you plan to keep or give an idea to your change rate. It's important to keep in mind that with V5, replica restore points were compressed in Veeam VRB format, while with V6, restore points are just VMware snapshots (Hyper-V still uses VRB rollbacks). These are obviously non-compressed and thus will require significantly more space.