Backup copy Question

[thatguycraig]

sorry for another post so soon! but heres a question....

again, doing a reverse incremental of 15 VM's on 3 hosts, all working well (side the mail server which ive posted about) but the question is -
whilst still in my trial of this, im doing a backup copy job to a 2TB SATA to emulate how I will take my backups off site, obviously being 2TB all my VM data wont fit on to that, so im using a selection of 3 servers from the main reverse incremental backup job. When I have checked the disk, I have one small .vbk file and two much larger .vib files

Just wondering why I have normal incremental backups instead of reverse ones? Is it because on the backup copy – I've only chose to take offsite 3 of the 11/12 backed up VM’s?

thanks for any input

[veremin]

It's by design. See explanations provided here. Thanks.

[foggy]

You can also review the corresponding user guide section for better understanding of the backup copying process.

[thatguycraig]

Thanks for your replies.

I however am still a little foggy after reading the info and the thread you posted.

My scenario will be this; I am taking reverse incremental jobs 3 times a day (not that this particularly matters, but seeing as its running pretty quick without disruption to the network, thought we would do away with shadow copies for in-hours restores and use Veeam for everything)

Over the weekend, I want to plug in a SATA disk locally on the server, and take offsite all of the previous weeks backup. Realistically I need to have the full backup, plus the increments for every day of the week (or in the above example, 3 increments per day, I just need to be able to revert to an earlier point in the week if necessary)

This copy job will run on the weekend and i will take it OFFSITE - i.e. Home with me, on Monday when I return. This is for disaster cases where the building has burned down and taken the onsite repository with it, I will have the SATA disk that i can restore from.

Is backup copy a good solution to do this? Or am I better running somthing else? As if you say "its by design" how do I show that I have a full, up to date backup, plus the ability to restore from one of the previous days of the week, if necessary?

many thanks

[foggy]

For that case you'd better use some file copy utility to copy the entire backup chain on the disk, since backup copy job will create a single restore point containing the latest VMs state.

[thatguycraig]

To use "some file copy utility" seems a little...... old school - as I would expect Veeam to handle this in some form or other. I think ive wrapped my head around the situation so i will see what tonights copy brings and come back with any problems!

[foggy]

To use "some file copy utility" seems a little...... old school - as I would expect Veeam to handle this in some form or other.

You can schedule file copy job for that.

[thatguycraig]

The problem with this is that you would have to copy the entire folder in order to run an automatic job?

for example: I will make a RAID array, that will amount to say, 8TB so that I can keep a large retention onsite.

For my offsite backup, I may only have a 4TB disk to take away with me, how can I Schedule it to just take away the most up to date increment automatically? I dont think you support this unfortunately :/ The only realistic way I can see me meeting the companies requirements of doing 3 reverse incremental backups a day, but just taking each says full backup offsite, is to run two separate backup jobs

[veremin]

Realistically I need to have the full backup, plus the increments for every day of the week (or in the above example, 3 increments per day, I just need to be able to revert to an earlier point in the week if necessary)

For my offsite backup, I may only have a 4TB disk to take away with me, how can I Schedule it to just take away the most up to date increment automatically?

The only realistic way I can see me meeting the companies requirements of doing 3 reverse incremental backups a day, but just taking each says full backup offsite, is to run two separate backup jobs

I'm slightly confused. Do you want to take offsite full backups, full backups plus all weekly increments or just the most up-to-date increment? Thanks.

[thatguycraig]

to take offsite - i can just have the most recent up to date FULL backup.

So i have reverse incremental running now 3 times a day, everyday.

On Monday i need to take offsite a full backup up until that point in time. on tuesday, a full backup of that point in time, etc. My idea, however, is to use the same disk each day, so increments are written to it rather than taking a FULL copy each and every day.

hope that make a little more sense! :/

[foggy]

I'm also confused by what you're trying to achieve. Previously you were talking about taking backups offsite on weekends only, so let's start from defining the actual required scenario.

Also, please note, that with reverse incremental mode full backup changes during each job run, while changes are offloaded to rollback incremental files.

[thatguycraig]

I've got my reverse incremental tied down. Thats all good, all understood and working well.

I'm probably confusing everyone with my terminology, as its really the way we describe it in house. Apologies for that. Let me try again.


So for our OFFSITE backups, I will be, or want to be, taking a copy of the backup away from the office with me on a removable SATA hard drive. Currently, a Full backup runs over the weekend, so when I come back into the office on Monday, I have a full backup ready to take offsite with me for disaster purposes. I am playing with doing this by Backup copy job at the moment. in my original post, I mentioned having the retention of 5 days, but you can ignore that, again, its just the way we talk in office that has confused my explanation. What I need, or would like to have, is a full functioning backup offsite with me each day. So the idea is to arrive in the office on Monday, and that disk will have a full backup on it that was taken over the weekend. I will then put in another disk, for the backup to write to on Monday night, so when I come in on Tuesday morning, I can swap the disks over, and take the SECOND disk offsite with me. That way I have a Full backup of the most recent data, should the whole place burn down before I get in on Wednesday morning.

Now, I presume that changing the disk over on Monday evening will mean that Veeam will write another FULL backup Monday night, as that disk will be empty? If that' the case, that's fine. The reason that I want to do this, is so the server is writing an offsite backup every night, but if the worst happens, a fire, a virus rips through the network etc, then I'm always going to have a Disk offsite with me that is a day old at the most.

After a LOT of reading and thinking about it today, I can see I have two options. Do as above with the backup copy job. I can set a retention of 5 days on this, so that the job creates a full job and 5 incrementals. This would be good for speed as Veeam is only writing a .vib each night and not a full .vbk. The trouble with this could be when I swap the disks over? Option 2 is I set the retention to 1 on the backup copy, and we get a full .vbk each night? not too much of an issue providing the copy is completed pretty quick and isnt interrupted when the ONSITE MAIN backup starts the next day.


Original question, is that the best way? Because the backup copy job is VERY SLOW in comparison to the actual backing up. Is this due to Veeam having to read the reverse incremental backup, and then having to create separate forward incremental files on the offsite disk? As far as I can see, my second option would in fact, be better, just starting a whole new backup job that targets the offsite disk, and running a reverse incremental on that with a 5 day retention.

I know there has been quite a few people on the forums and other sites that have this issue / concern. I think the backup copy job would be good for smaller backups, but not for us doing 1 main full backup of all 16 VM's.

Lastly, the file copy would be great - I did a test on a 95GB .vib file and it took 15 mins - but you would need to manually do that each and every day, or target the full folder, which means that you couldnt keep a big retention on disk, without moving some files out, again manually.

[foggy]

So, basically, you need to copy the latest full daily on a rotated drive. I would suggest to use backup copy job with a daily interval for that, which will copy VMs latest state right after the source backup job completes. Provided you configure repository cleanup, you can set any retention for the backup copy, since it will never go far enough to create even the first increment.

Alternatively, you can use any third-party utility to copy the latest full backup to the newly inserted drive.

[thatguycraig]

So, basically, you need to copy the latest full daily on a rotated drive. I would suggest to use backup copy job with a daily interval for that, which will copy VMs latest state right after the source backup job completes. Provided you configure repository cleanup

i think ive sort of grasped it. No great Information and guides knocking about for this but a full day worth of reading and I think it will work as I expect, rotating two disks.

you can set any retention for the backup copy, since it will never go far enough to create even the first increment.

i dont get this - as surely setting a 2 day retention, for example, the disk would have a full backup file and then one increment?


Let me point out one BIG flaw in this for Veeam that someone should pass on. The backup copy job is continuous and runs after the set time, AND after each and every increment job. So for us, the backup copy is running 3 times a day. I can set the schedule to not allow this, but if the job runs over for some reason and the copy isnt finished, but the schedule is set to 'disallow' then my copy ends and errors out, waits for the next interval to begin.

it would be highly advantageous, and im sure not just for me, to be able to tick a box that says "copy once" so the copy runs one time, once a day

[foggy]

i dont get this - as surely setting a 2 day retention, for example, the disk would have a full backup file and then one increment?

Unless you change the drive. In a daily rotation scenario and daily sync interval increments will never be created.

Let me point out one BIG flaw in this for Veeam that someone should pass on. The backup copy job is continuous and runs after the set time, AND after each and every increment job.

That's not correct. Backup copy job copies one VMs state within each sync interval, so regardless of the source backup job schedule, you can set your backup copy to copy once a day.