Host-based backup of VMware vSphere VMs.
Post Reply
singy2002
Service Provider
Posts: 66
Liked: 1 time
Joined: Mar 22, 2010 8:43 am
Full Name: Andrew Singleton
Contact:

Values in replica jobs

Post by singy2002 »

Hi,
I am trying to understand exactly what is slowing down some replica jobs, before we upgraded to V6, we had a job for one managed customer which replicated 5 VM's in about 3 hours, since upgrading to 6 it now runs for 8-9 Hours.

I am looking through the reports and see each machine taking certain amounts of time and each step of it processing different disks. But I dont understand the "read" value, is this what has changed on the disk?

Example:
Hard Disk 2 (355GB) 47.2GB read at 4MB/s (CBT)

This server in particular is taking along time, over 3 hours alone.

For info the link in use between source and destination is a 100MB LL.
foggy
Veeam Software
Posts: 21070
Liked: 2115 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: Values in replica jobs

Post by foggy »

Yes, this is a real speed at which the changed blocks of the disk were read. Btw, what is your replication architecture? Do you have proxies set up both at the source and target locations? Do you store replica metadata on repository located in the source site?
singy2002
Service Provider
Posts: 66
Liked: 1 time
Joined: Mar 22, 2010 8:43 am
Full Name: Andrew Singleton
Contact:

Re: Values in replica jobs

Post by singy2002 »

Hi foggy,

Yes we have a proxy set at both source and destination, we throttle the traffic during th day so as not to impact the other traffic on the LES.

as for replica metadata, we store it on the target host.

So for the example I have given, does that suggest a daily rate of change of 47GB (meaning that amount is transfered over the line?

Regards
Andy
foggy
Veeam Software
Posts: 21070
Liked: 2115 times
Joined: Jul 11, 2011 10:22 am
Full Name: Alexander Fogelson
Contact:

Re: Values in replica jobs

Post by foggy »

singy2002 wrote:as for replica metadata, we store it on the target host.
You should definitely change the repository storing replica metadata. As explained in the UI when creating replica job, backup repository should be located closer to the source. Placing the metadata in the remote site slows down the replication process significantly (depending on the metadata size and the link speed) as the source agent pulls it over WAN.
singy2002 wrote:So for the example I have given, does that suggest a daily rate of change of 47GB (meaning that amount is transfered over the line?
That number indicates the amount of data that has been read from the source VM disk (a daily rate of change, if CBT is enabled). The actually transferred amount of data (blocks which are different between the source and the target disks) is displayed to the right from the "Transferred" label in the "Data" section at the top of the "Realtime Statistics" window.
Post Reply

Who is online

Users browsing this forum: aQuestionmark and 75 guests