raid-5 volunerable to shutdown/reboots?

[theinv]

I have the Areca 5020 raid box with 4 x 1TB Samsung EcoGreen drives.

It has happened a few times that after I shutdown/reboot the raid box, I got the device failure error and the volume got degrade. Sometimes, if I shutdown and restart again, the error is gone and all disks are functioning. But the very last time, one of the member disk got "device failure" again and rebooting the raid box no longer helps. Then I deleted the raid volume set, deleted the raid set, recreate raid set, recreate the volume set. Basically, I restart from scratch, and all disks now seems healthy. Of course, I lost all my data. I am so afraid of rebooting the box now because every time I reboot, I will get some kind of errors like "device failure", "incomplete raid set" or "volume degraded".

Is raid5 only designed to run 24x7? I don't want to have my box running when I am not using my computer. If raid5 cannot survive reboots, how can it claim to be more reliable than raid0, raid1 or single disk setup? To me, raid5 is the worst raid of all. Other raid setups never have this problem.

 

[Lazn_Work]

it isn't RAID 5 that is causing the problem. I have run RAID 5 for decades on servers without these errors on hundreds or thousands of reboots.

However RAID of any type does add complication and additional points of failure to a system, and usually it's purpose is to increase uptime (that is keep the system running when a drive fails) or performance and not to keep data safe (that is your backup's job)

It sounds to me like your problem is an incompatibility between your Areca and the Samsung drives.

 

[geiger]

Do you have staggered spinup configured? If the issue only occurs at reboot is sounds like one of the drives is not fully initialized before the array does a health check.

With my 3ware card it will generally kick the drive out as a JBOD, I have to delete the JBOD volume, then do an online rebuild to get the drive back into the raid volume.