1 Rookie
•
99 Posts
0
2588
7910T freezes with "Shutting down filesystem" error
I installed Centos 7 on a 7910T. When it boots it shuts down after a few minutes with the errors below. I am not sure how to debug whether it is a disk issue or raid controller or something else?
I tried running the ePSA diagnostics but it hangs after some time on some WCMATS test. See picture beneath this one.
thisisalloneword
1 Rookie
1 Rookie
•
99 Posts
0
March 9th, 2021 13:00
Great, @bmcowboy After removing the broken raid card I was sucessfuly able to install centos 7 on the machine! I had to press Ctrl-C on bootup to first configure the four drives. I am not sure how those settings work. I think I just did Raid 0, so that all four 1.8 TB drives combine space into one large 7.4 TB drive . I am guessing this is risky if one drive fails then the whole system fails?
When I removed the Raid card I noticed one screw attaching the aluminium heatsink was missing. So perhaps that chip overheated or something..
bradthetechnut
6 Professor
6 Professor
•
8.2K Posts
0
March 9th, 2021 14:00
I almost forgot... The system does fail if one drive fails. If you Google it, you'll find even more info.
bradthetechnut
6 Professor
6 Professor
•
8.2K Posts
0
March 9th, 2021 14:00
Great to hear you got it working!! What an experience you had.
The only thing I was going to mention is if the screw holding the raid card is holding anything else, be sure to put it back. I wanted to let @bmcowboy answer first. Or may be you can use that screw for the missing heatsink screw if it's the same size and thread.
@bmcowboy - The only thing I have to mention is the "blue caps" you referenced are tabs. You did a great job helping @thisisalloneword and I'm glad you came onboard.
mazzinia_
4 Operator
4 Operator
•
1.1K Posts
0
March 9th, 2021 15:00
If you want some security
Raid 1 : half disks will be used , and with 2 x raid 1 you get hdd1 (and its security copy) and hdd2 (and its security copy). You can survive losing 1 to at most 2 drives.
Raid 10 : half the disks will be used (hdd1+2) to create a single hdd, other half will be a copy that you cannot use. You can survive losing 1 to at most 2 drives.
Raid 5 : 3 disks for you, to create a single hdd , 1 for safety. You survive losing 1 drive
If you want more space available to you, go with Raid 5
bmcowboy
573 Posts
1
March 9th, 2021 18:00
Hi @thisisalloneword ,
Glad to know that you finally get your hard drives running. And thank you @bradthetechnut @mazzinia for explanation on RAID settings. Yes, RAID0 is risky especially when you target drive group for data storage without backup. People often use RAID0 on OS only while it could be reinstall at anytime. However, data lose due to hard drives failure on RAID0 is irrecoverable in most cases. That's why people tend to use RAID10 (= raid1 mirror + raid0 stripe) or even RAID5/6 for data storage.
While your original question related with "dev sda sector" error logs, suggest to find some HDD health check tools and perform thorough scan on all four hard drives to make sure that they are in good condition.
Further suggestion: Instead of installing OS on RAID volume built by mechanical hard drives, better grab a 2.5" SSD for OS and left RAID for data storage only. A single SSD provide higher performance and stability when compare with RAID volume. Check out the picture I provided before regarding SAS0 port. Number 33 is the two spare SATA ports on motherboard. You can use one of them for SSD connection and place the drive in the spare room right above the 4 3.5" hard drive bays.
bradthetechnut
6 Professor
6 Professor
•
8.2K Posts
0
March 9th, 2021 19:00
@thisisalloneword ,
I do this from my phone and can't see photos while signed in. So I didn't see the photo of your raid card right away and didn't know you were referring to its heatsink. I first thought you were referring to CPU heatsink and thought that's where you had a screw missing.
Neat article on raid if you want: https://www.prepressure.com/library/technology/raid