SoftRAID Support Forum
OWC 5.7.2 driver anomaly? - Printable Version

+- SoftRAID Support Forum (https://www.softraid.com/forum)
+-- Forum: Solving Problems (https://www.softraid.com/forum/forumdisplay.php?fid=6)
+--- Forum: Mojave Issues (https://www.softraid.com/forum/forumdisplay.php?fid=19)
+--- Thread: OWC 5.7.2 driver anomaly? (/showthread.php?tid=1019)



OWC 5.7.2 driver anomaly? - KMarc83 - 12-13-2018

I have been running a Thunderbay 4 on Thunderbolt 3 with 4 8TB seagate ironwolf drives in RAID 5 under High Sierra (SoftRAID XT 5.7 and 5.7.2) for months with no problems of any kind. The machine is a 2018 iMac 27" Retina with 64gb memory.

The raid was (obviously) a 24TB raid drive but I divided it into two 12TB volumes, both of which mounted automatically. Again, I had never experienced any issue with my RAID.

I've been waiting to upgrade to Mojave and for some idiotic reason I did so a couple of nights ago. I started the upgrade and went to bed.

In the morning, one of my two RAID 5 volumes was mounted grayed-out with a red exclamation point, and the other was not mounted at all. Both should have mounted automatically and always did before.

The SoftRAID 5.7.2 application said something about an I/O error on one of the four drives, but also said that drive's SMART status was OK and on another line said "0 errors". It was simply too much of a coincidence that the drive would just happen to go bad during the OS upgrade. I took the "bad" drive out and reseated it. Same problem. I put in an identical spare 8TB drive and started a rebuild. Same thing, the rebuild would get to a certain point and then hang, and the volume would refuse to mount.

The Mac OS disk utility would successfully run First Aid on BOTH RAID partitions, but would never mount one of them, and the SoftRAID utility would always say "waiting for mount".

Since I have both RAID volumes (mostly) backed up, I decided what the hell, and used SoftRAID 5.7.2 to destroy the RAID by initializing all 4 drives, and then rebuilt the RAID and both of its partitions in exactly the same configuration as before, but now of course empty.

I am now a couple of hours into the all-night process of restoring from the backups. No problems at all. The restore is chugging right along and the utility shows no errors on any drive or on either RAID volume.

Obviously this is a software problem, an incompatibility with SoftRAID and Mojave, that will probably never be found, but I thought I might as well report it anyway.


RE: OWC 5.7.2 driver anomaly? - SoftRAID Support - 12-14-2018

Every major OS release we get s few cases like this. Our theory is that the OS upgrades various things on all the volumes, then shuts down before all I/O's can be completed on all the disks.

this can cause minor volume directory damage. It keeps Disk Warrior in business, as HFS volumes then either cannot mount, or mount "read only".

the repair in Disk Utility generally is not useful, only Disk Warrior, or as you did, restore from backup.

We had added a subtle feature in 5.6.7 to attempt to reduce these incidents, but it is not a real solution. But we started seeing this with SSD startup volumes increase in frequency, which led to our theory.

We expect that APFS volumes will prevent this, as directories have a more robust structure. SoftRAID will support APFS next year, we expect in the spring.