I have about 600,000 pdf files that exist in 8 vhd files, which i am mounting (right click -> mount) and copying to a storage spaces direct cluster disk as individual files. When using robocopy with more than 64 threads, eventually the cluster node barfs with "Insufficient resources exist", or similar. Once this error is encountered, i have to restart the node, because you can't access the volume after that. So i just retried the operation with less threads. OK.
Now I am trying out hyperoo, a backup program, and when i try to back up this volume, it fails with these messages in it's log:
[12/12/2016 4:14:28 PM] - *ERROR*- Unable to access the file "arch\2016trim\30-16\25\25-30-16-98396-003-3060.pdf" - System.IO.IOException - Error accessing file C:\HrMnt\3a6\VOLUME1\arch\2016trim\30-16\25\25-30-16-98396-003-3060.pdf to get file
information - Insufficient system resources exist to complete the requested service. (Exception from HRESULT: 0x800705AA)
[12/12/2016 4:14:28 PM] - *ERROR*- Unable to access the file "arch\2016trim\30-16\25\25-30-16-98396-003-3070.pdf" - System.IO.IOException - Error accessing file C:\HrMnt\3a6\VOLUME1\arch\2016trim\30-16\25\25-30-16-98396-003-3070.pdf to get file information
- Insufficient system resources exist to complete the requested service. (Exception from HRESULT: 0x800705AA)
(repeats untill i kill the job)
This program uses VSS snapshots.
So I try to run windows server backup on the volume, and it fails similarly:
Error in backup of C:\ClusterStorage\Volume1\ during read: Error [0x800705aa] Insufficient system resources exist to complete the requested service.
Error in backup of C:\ClusterStorage\Volume1\ during read: Error [0x800705aa] Insufficient system resources exist to complete the requested service.
ad nauseum.
Again, after these messages show up, pretty much have to restart the cluster node.
My guess would be some type of pool exhaustion. Anyone out there know how to confirm, or do debug this further?