Split backup jobs due to DR copy job taking too long

[masogeo]

Hi,

We have a daily backup that runs to local storage within our DC of approximately 40 VMs (about 10TB processed per night). Once weekly we have a backup copy job on a GFS rotation which goes to offsite DR hosted in AWS via a Virtual Tape appliance.

This works well, however the volume of data has grown to the extent that the AWS backup copy job takes too long to run, meaning that the source job conflicts with it. For the time being we have worked around the issue by changing the schedule of the daily job not to run at weekends, whereby the copy job completes without errors. However the problem is that we need to run the source job 7 days per week.

What is the best way of splitting these jobs, so that the backup copy can successfully complete, but without impacting the daily job? Should I just split the copy job, and specify which VMs to back up in each of the copies?

Thanks

G

[bdufour]

ive ran into similar issues, basically you need a faster circuit to your dr site or a wan accelerator (depending on your circuit size) if your environment is growing to the point where managing this is becoming near impossible, and trust it will get to the point where its impossible. talk to management.

[masogeo]

Trouble is, the DR site is AWS, and we have more than enough bandwidth - it just seems that for whatever reason, the AWS tape job runs at 25Mb/s, and so takes an age to complete.

So I guess there are two fixes - either speed up the tape job, or split the backup copy job.

FWIW Veeam support already told me that I should split the copy job, but I was interested to hear from other members of the community on what they'd experienced.

[bdufour]

i see, what are your bottlenecks looking like in that job?