-
- Influencer
- Posts: 10
- Liked: never
- Joined: Aug 12, 2009 5:05 pm
- Full Name: Andy
- Contact:
removing vm backups and backup resiliency
Running 4.1.1 in a vm on 2008R2 x64, 100% NFS datastores (10gbe):
Have jobs configured basically as "host" (items job contains is only vcenter host, which means anything there gets backed up). Someone creates a "test" vm and "accidentally" leaves it associated with the ESX host and it gets backed up. 20 days later I still see restore points for that VM - tried setting retention to 1 and it didn't disappear. Other than completely delete the job.. or even the backup files, triggering a full backup again, is there any other way to age out vm backups for vm's that no longer exist?
Second question (which is another source of frustration when folks like moving VM's around):
is it possible to have many unique jobs all pointing at the same dedupe store (vrb file on cifs)? Also noticed that when doing data-center item backups and a host moves and falls under a different job, that the whole thing is taken again; also leaving the old backup to hang out on the old store, until I redo the backup job completely.
Best practice to have backups targetted to VM, ESX Host, Data Center, or vcenter? I don't want to reconfig jobs every time someone decides to do a vmotion
if I need to clarify anything, let me know, or if this is best suited as a support case - we have that too
Im open to selecting individual VM's in backup jobs.. but that still doesn't solve the host vmotion thing unless I can have multi-job per dedupe store right?
TIA!
Have jobs configured basically as "host" (items job contains is only vcenter host, which means anything there gets backed up). Someone creates a "test" vm and "accidentally" leaves it associated with the ESX host and it gets backed up. 20 days later I still see restore points for that VM - tried setting retention to 1 and it didn't disappear. Other than completely delete the job.. or even the backup files, triggering a full backup again, is there any other way to age out vm backups for vm's that no longer exist?
Second question (which is another source of frustration when folks like moving VM's around):
is it possible to have many unique jobs all pointing at the same dedupe store (vrb file on cifs)? Also noticed that when doing data-center item backups and a host moves and falls under a different job, that the whole thing is taken again; also leaving the old backup to hang out on the old store, until I redo the backup job completely.
Best practice to have backups targetted to VM, ESX Host, Data Center, or vcenter? I don't want to reconfig jobs every time someone decides to do a vmotion
if I need to clarify anything, let me know, or if this is best suited as a support case - we have that too
Im open to selecting individual VM's in backup jobs.. but that still doesn't solve the host vmotion thing unless I can have multi-job per dedupe store right?
TIA!
-
- Chief Product Officer
- Posts: 31793
- Liked: 7295 times
- Joined: Jan 01, 2006 1:01 am
- Location: Baar, Switzerland
- Contact:
Re: removing vm backups and backup resiliency
Hello Andy,
1. In the current version, the only way to remove VM from backup file is to initiate Full Backup; however in the next release VM's will age out automatically according to the retention policy.
2. No, it is not possible - we are not doing global dedupe.
3. Recommended way to target backup jobs in the environment with VMotion in use is using VM folders or clusters. VM folders is what majority of customers seem to be using. Adding individual VMs is generally not recommend, but can help in some cases when you want to have more control on what is included in the job. VMotion should not be causing any issues to jobs configured with individual VMs added as long as you have configured your have vCenter server added to Veeam Backup Servers tree, as opposed to individual ESX hosts.
Thanks!
1. In the current version, the only way to remove VM from backup file is to initiate Full Backup; however in the next release VM's will age out automatically according to the retention policy.
2. No, it is not possible - we are not doing global dedupe.
3. Recommended way to target backup jobs in the environment with VMotion in use is using VM folders or clusters. VM folders is what majority of customers seem to be using. Adding individual VMs is generally not recommend, but can help in some cases when you want to have more control on what is included in the job. VMotion should not be causing any issues to jobs configured with individual VMs added as long as you have configured your have vCenter server added to Veeam Backup Servers tree, as opposed to individual ESX hosts.
Thanks!
-
- Influencer
- Posts: 10
- Liked: never
- Joined: Aug 12, 2009 5:05 pm
- Full Name: Andy
- Contact:
Re: removing vm backups and backup resiliency
1. OK thats what I thought - I don't know if I want automagic age out though, perhaps a manual dedupe-store pruning tool? If I set retention to 6 and 5 failures happen, I don't want 1 backup on disk. Also I could see that if a vm gets deleted and 2 weeks later someone is asking for it. But I also don't want to have to fully copy 6tb just to age out an old 100gb vm from the backup.
2. OK, now to #3 then
3. excellent - we didnt know veeam could see folders as entities. This really gives me some good ideas!
As a side note - not many of your competitors products are even mature enough to have these dilemna Keep up the good work!
2. OK, now to #3 then
3. excellent - we didnt know veeam could see folders as entities. This really gives me some good ideas!
As a side note - not many of your competitors products are even mature enough to have these dilemna Keep up the good work!
-
- Chief Product Officer
- Posts: 31793
- Liked: 7295 times
- Joined: Jan 01, 2006 1:01 am
- Location: Baar, Switzerland
- Contact:
Re: removing vm backups and backup resiliency
We found that almost all customers keep a few weeks of backups on disk (for short term restores), but also take full backups to tape or removable hard drive or other media every week or month to keep for long term restores. So autopruning is designed with this in mind. Do you only keep backup files on disk and don't copy them periodically offsite?alubel wrote:Also I could see that if a vm gets deleted and 2 weeks later someone is asking for it.
That's right, it may not obvious at a first sight, but you can switch views when adding objects to the job between Hosts and Clusters view, and VMs and Templates view which shows VM folders.alubel wrote:we didnt know veeam could see folders as entities
Thank you for your kind words
-
- Influencer
- Posts: 10
- Liked: never
- Joined: Aug 12, 2009 5:05 pm
- Full Name: Andy
- Contact:
Re: removing vm backups and backup resiliency
Yeah we do go offsite every 14 days and my retention is tailored to that, so you are right. I just haven't tried importing one of those yet.
side question
"Linux Host" vs. CIFS backup targets. Is it just really SSH (encryption overhead) vs samba (veeam is the only reason samba is running)?
side question
"Linux Host" vs. CIFS backup targets. Is it just really SSH (encryption overhead) vs samba (veeam is the only reason samba is running)?
-
- Chief Product Officer
- Posts: 31793
- Liked: 7295 times
- Joined: Jan 01, 2006 1:01 am
- Location: Baar, Switzerland
- Contact:
Re: removing vm backups and backup resiliency
We are not use SSH as data transfer channel with Linux targets. We are using proprietary protocol (direct socket to socket connection between our agents) specifically to avoid SSH overhead and to be able to use full network capacity. So if you are backing up to the Linux host, there is really no reason to run samba for CIFS there.
Who is online
Users browsing this forum: Bing [Bot], Google [Bot] and 83 guests