SCDPM Out of Disk Space

[Gongalong]

Hi folks,

 

I'm starting to get insufficient storage space on our DPM (2012) server.

 

I've reduced the retention range on some of the jobs. How good is DPM at updating its volume allocations? i.e. will it reduce the volume allocation if it needs less space.

 

If not, is there any routine I can run to get it to reassess/compact the data its using?

 

TIA

[cscott]

DPM estimates the amount of disk space it needs for replica and recovery point volumes estimated on what it thinks the growth rate will be, as a result I've found it can often allocate that volumes that are too large, especially if you enabled auto-grow - you can shrink the volumes though to recover this over estimated space.

 

You could try right clicking on one of your protection groups and selecting... a dialog box will open showing how much space is allocated to each replica & recovery point volume, there should be a "Shrink" link next to each recovery point volume, if you click the link DPM will recalculate the space needed for that recovery point volume and if possible reduce the volume.

 

Also have you got anyting under the "inactive protection for previously protected data" protection group? depending on how you removed this items from your protection group they could still be using up disk space - tends to happen if you removed them via modifying the protection group as doesn't give you option of deleting the replicas from disk.

[twin--turbo]

What exactly is the message your getting?

 

Rob

[Gongalong]

cscott: I've run through every group, albeit most won't shrink ("data at the end of the volume"). I suspect I'll need to leave it over the weekend post-shrink to see if that has helped, at least short term.

 

There are no inactive groups unfortunately. Of course it's somewhat impossible to say what volumes are being used...

 

twin--turbo: A couple of different errors:

 

DPM has run out of free space on the recovery point volume and will fail synchronization for Server XYZ in order to prevent existing recovery points from being deleted.

Used disk space on replica volume exceeds threshold of 90%.

[cscott]

Just read your error message - it's a complete different issue from what I assumed. I thought you meant your physical disk was full, this just sounds like DPM hasn't been able to autogrow a volume.

 

This could be cause by the disk being full so probably worth checking you've got space on your disk by going to the "Management" view and selecting the "Disk" link, check how much free space the disk has.

 

Assuming you've got free space on your disk the error message your getting just indicates that one of the volumes on your disk is full and DPM couldn't autogrow it. Go to your protection groups and locate Server XYZ and select the resource that it can't back up (will be shown in red). Right click and select "Modify volumes", same window as before but it will only show volumes for this server, go to the replica volumne colomn and increase the size of the replica volume by at least 25%.

[Gongalong]

Ah, well if I tried to grow the volume it would say there was insufficient local disk space.

 

Shrink has had some effect because I can now grow the two that are reporting with errors.

 

I'm sure medium/long term there will still be an issue, because our file servers inevitably just increase in disk space usage.

[cscott]

yeah, sounds like you to add some disks to your DPM server!

[Gongalong]

We have a Buffalo Terastation, and I have a horrible feeling it's already full... (a system I inherited)

[Gongalong]

Lots more errors this morning, so clearly the limit is definitely reached. This is the error I get when I try to modify a disk allocation:

 

There is insufficient space on the storage pool disks to allocate the replica and recovery point volumes

 

Add more disks to the storage pool by using the Disks tab in the Management task area, or reduce the specified allocations

 

ID: 171

[cscott]

Yep sounds like the terrastations full...

[rh91uk]

Yes, you need to add more disks/remove things - how about offloading some of it to tape maybe?

[jofloz]

Gongalong how well does terastation work with DPM? Which model do you have?

[Gongalong]

Apologies for the lack of follow up.

 

We had a (Buffalo) TeraStation III iSCSI Rackmount TS RIXL. Fortuitously or not it failed, with what appears to be a common fault, so we switched to an HP Storeeasy with lots more space. I can't recommend the Buffalo because it failed less than 2 years from purchase.