Hi everyone,
I have experienced some problems with iSCSI on my N5200PRO (5x1TB RAID5) which are concerning to me. The unit works fine for some time, but when it is put under a bit of load (backing up 800GB) the performance suddenly degrades and sometimes a partition of the iSCSI target stops working altogether after a later shutdown/restart. The other partitions continue to function.
I mounted the iSCSI target and HFS+ partitioned it into 5 using Disk Utility totaling 2.69TB. The smaller partitions I use for Time Machine backups for each for the Macs I own. The other partition is 2TB for my data. It is this partition which stops working and appears grey in Disk Utility after storing 1.2TB on it. If I try to mount the partition it reports that it is unable to mount and I should "Try running First Aid on the disk..."
Verifying the disk reports:-
Checking catalog file.
Invalid sibling link
The volume was found corrupt and needs to be repaired.
The Repair Disk function reports
Rebuilding catalog B-Tree
Disk Utility was unable to repair this disk....
I have lost data due to this problem which has occurred on 2 different RAID builds (previously RAID6). I am very dissatisfied with this. I took great care to shut down cleanly at all times with no power failure and the Storage hardware is always reported as 100% healthy.
This can also be achieved if you have two separate Mac clients using the same iSCSI target and write data with one Mac at the same time as making changes to a different partition e.g. renaming. This of course un-mounts then mounts one of the partitions. There is no protection in avoiding this obviously dangerous operation! Do enterprise solutions deal with this sort of issue?
Has anyone else had similar issues with iSCSI and GlobalSAN initiator?
I would be most grateful if you could advise me about what to do to fix this problem. I think I might have to buy an enterprise quality piece of hardware and stop pretending the Thecus products are ever gonna be rock solid.
Kind regards,
Steve.