Backup every 12 hours, ONLY ON WEEKDAYS?

[briggs]

Hi, Basically, is backing up every 12 hours still possible without running on weekends?
We are retaining 3 months worth of backups of our main critical file server, but the storage required at the repository is becoming too difficult to manage. We have 8TB available to us, but already it's full after 93 days!
Can I stop my 12 hourly backup job running on a Sat and Sun at all? This will help only a little bit, as they're incremental anyway, but every little helps.

I could experiment with the compression and optimization settings I suppose, but right now, I'd rather not experiment on a critical backup. I'm using "Optimal - Recommended" compression and optimise for LAN Target. Would changing these make much of a difference, without crippling the backup processing times?

I have to get the size of these things down a bit, I hate to keep comparing products, but Arcserve D2D used to cope with 6 months worth of restore points without running out of space!

Thanks, Alan

[veremin]

Hi, Basically, is backing up every 12 hours still possible without running on weekends?

Yep, you can. Just go to the Job settings -> Schedule -> periodically every -> Schedule -> and select required interval (Saturday, Sunday) during which the job execution should be prohibited.

Thanks.

[briggs]

Boy do I feel stupid I can't believe I missed the fact you can de-select days from the grid! Thanks for pointing out what should of been obvious! I don't suppose you have any comments on the compression/optimization at all?

[veremin]

Providing that corresponding option has answered your requirements, I've assumed that you’re not that interested in compressions settings any longer.

Moreover, it’s rather hard to make any speculations regarding this or that particular setting without performing actual tests, since it’s always a matter of certain option dependency on numerous factors that can be hardly taken into account via forum correspondence.

Of course, it’s likely that extreme compression will put additional pressure on your backup infrastructure, that, indeed, might lead to decreased performance. However, running several tests should this question for sure.

Thanks.

[foggy]

Seems that you already have optimal settings in terms of compression and deduplication. Setting compression to Extreme level will literally put extremely high load on your proxy servers, however will not provide big savings. Extreme level allows to save about 2%-3% of space at the expense of double CPU load.

What you could additionally do to decrease backup storage space usage down further, is to group similar VMs together in one job to get better dedupe among them (unless you've already done that). Another point to look at is your guests from perspective of how many changes do they produce between job cycles. For example, such activities as scheduled defragmentation and antivirus scans perform a lot of changes to the disks, which result in large increments during backups. And, of course, reverse incremental mode is much more space-friendly that the forward one, although puts higher load on the target storage itself.

[yizhar]

Hi.

We are retaining 3 months worth of backups of our main critical file server, but the storage required at the repository is becoming too difficult to manage. We have 8TB available to us, but already it's full after 93 days!

Actually, disabling backups during weekend will not make a change nor solve the issue.
Compression and deduplication changes won't do it either.

I suggest changing the design a bit.

Keeping a daily chain for 90 days (twice a day) isn't a good idea IMHO.
I think that you should divide the backup jobs for short and longer term retention.

For example:

Configure a daily backup job, keeping about 10 last days only (which means 20 restore points in your case of 2 backups a day),
using either forward or reverse incremental depending on other factors and your preferences.

Configure another job which will run weekly (once a week) using reverse incremental, and keep last 20 restore points (about 4 month worth).

This can be done for all VMs, or specifically for the file server, depending on your needs.

Yet another possible design and approach:

1. Install windows 2012 as the backup server instead of windows 2008.
2. Divide the 8tb disk to 2 volumes for short and long retention.
Enable deduplication on the long term retention drive (only).
3. Configure daily job as above, using the repository on short term drive.
4. Use the new "backup copy job" in upcoming V7, with GFS rotation,
and configure it to keep weekly backups for X weeks, on the long term drive with global deduplication from OS.

Yizhar

[briggs]

Thank you Yizhar, some great tips that I may well have to look into! Thanks again. Alan