BackUp device for local or colo storage
-
@DustinB3403 said:
@Dashrender My concern is only having 1 Backup of my vDISK from Xen might not be enough for recovery.
Should something happen to the file the guest VM is hosed (for recoverability). Having multiple backups seems more logical. But needs to be planned.
This doesn't seem to be a concern for Veeam or Unitrends - both can use incrementals forever (though I think they might suggest doing a full once a month).
-
@Dashrender said:
@DustinB3403 said:
@Dashrender My concern is only having 1 Backup of my vDISK from Xen might not be enough for recovery.
Should something happen to the file the guest VM is hosed (for recoverability). Having multiple backups seems more logical. But needs to be planned.
This doesn't seem to be a concern for Veeam or Unitrends - both can use incrementals forever (though I think they might suggest doing a full once a month).
Full monthly to local tape might work. LTO 5 at 3TB would mean taking home 3 tapes a month. Not bad..
backup to disk first on the R510 and to tape.
-
@DustinB3403 said:
@Dashrender said:
@DustinB3403 said:
@Dashrender My concern is only having 1 Backup of my vDISK from Xen might not be enough for recovery.
Should something happen to the file the guest VM is hosed (for recoverability). Having multiple backups seems more logical. But needs to be planned.
This doesn't seem to be a concern for Veeam or Unitrends - both can use incrementals forever (though I think they might suggest doing a full once a month).
Full monthly to local tape might work. LTO 5 at 3TB would mean taking home 3 tapes a month. Not bad..
backup to disk first on the R510 and to tape.
Depending on the dedup, you might be able to get that to two or even one tape.
-
Veeam Endpoint recovery or Veeam B & R would handle the full backups as well as the incrementals. Veeam B & R may not work with Xenserver. I'm not sure about that.
Veeam EPR is installed per VM and not centrally managed.
The 24 TB total he is calculating appears to be what he would need if he is doing disk to disk backups.
-
@Dashrender said:
@scottalanmiller said:
Can you do incremental forever to tape? I don't believe so. He would need to hit the WAN. 8TB a day is pretty ridiculous to go over a WAN.
Why would you need to send 8 TB a day via the WAN? He'd only send the changes daily via the WAN. And AppAssure supports that type of replication.
This is why knowing the daily delta would be important.
Didn't he say the incrementals needed 8TB today?
-
@dafyre said:
Veeam Endpoint recovery or Veeam B & R would handle the full backups as well as the incrementals. Veeam B & R may not work with Xenserver. I'm not sure about that.
Definitely does not.
-
@scottalanmiller said:
@Dashrender said:
@scottalanmiller said:
Can you do incremental forever to tape? I don't believe so. He would need to hit the WAN. 8TB a day is pretty ridiculous to go over a WAN.
Why would you need to send 8 TB a day via the WAN? He'd only send the changes daily via the WAN. And AppAssure supports that type of replication.
This is why knowing the daily delta would be important.
Didn't he say the incrementals needed 8TB today?
No 8 TB was to perform weekly fulls at the current infrastructure size.
Incremental changes are likely > 20 Gb/day
-
Which if I ignore performing weekly fulls and instead perform them monthly to tape. I'll have to start looking into LTO tech has I have zero experience with it.
Edit: Well I should probably look into LTO tape anyways...
-
I think the solution really is to perform monthly fulls to local disk then to tape and take the tapes home.
Perform incremental backups using storage craft to on-site / offsite daily and dump the old versions at 72 hours. Cycling the tapes every 72 hours (3 days)
Would be easier to just keep it for 5 days though.... This way we have to swap less tape around...
-
@DustinB3403 said:
No 8 TB was to perform weekly fulls at the current infrastructure size.
Wait, if a full is 8TB, where did I get 22TB from?
-
@DustinB3403 said:
Incremental changes are likely > 20 Gb/day
Oh that's nothing. Just seed a backup device and set it remote.
-
@scottalanmiller said:
@DustinB3403 said:
No 8 TB was to perform weekly fulls at the current infrastructure size.
Wait, if a full is 8TB, where did I get 22TB from?
Full Backup retention for an entire month.
-
@DustinB3403 said:
@scottalanmiller said:
@DustinB3403 said:
No 8 TB was to perform weekly fulls at the current infrastructure size.
Wait, if a full is 8TB, where did I get 22TB from?
Full Backup retention for an entire month.
That would be 32TB then.
-
Actually it's technically a 7 day swap cycle now for what we have.
Might have to keep the same process if I want to reduce IT's labor on the backup process. (Weekends count to computers)
-
@scottalanmiller said:
@DustinB3403 said:
@scottalanmiller said:
@DustinB3403 said:
No 8 TB was to perform weekly fulls at the current infrastructure size.
Wait, if a full is 8TB, where did I get 22TB from?
Full Backup retention for an entire month.
That would be 32TB then.
8 TB is with growth... currently he's at just over 6 TB.
so 24 TB for a month -
@scottalanmiller Only if we we're fully using all 8TB.
We have free space on the systems now and are only using approximately 6TB of total 8TB capacity in the infrastructure.
-
I'm really lost - why are you worrying about full backups when you have a StorageCraft backup solution?
-
@Dashrender So if something should happen to the vDISK on XS. That we have a way to simply import a 2TB file to the server, and turn it on.
Trying to K.I.S.S
-
@DustinB3403 said:
@Dashrender So if something should happen to the vDISK on XS. That we have a way to simply import a 2TB file to the server, and turn it on.
Trying to K.I.S.S
Can't you get that from the StorageCraft system?
-
Not effectively (at least with what we have).
Fulls currently take over 24 hours to generate when their needed. So fulls are almost never taken.
But our system partitions are ignored, besides 2 servers.
I want to include everything in the backup process. So everything is protected.
