Ubuntu 20 Server and Raid issue

[mdrabble]

I thought I had managed to install Ubuntu 20 on an repurposed Windows Server which use to hold VM data, so I could use it to hold Veeam immutable backups.

 

I wiped the array and reconfigured it so that the existing arrays are as follows.

 

Server has a MegaRaid (Avgo) Raid Controller.

 

8 x 120GB (Raid 0) SSD Caching Arrary

16 900TB (Raid 6) Data Arrary.

 

and 1 x SSD Intel onboard controller for OS

 

Installed Ubuntu 20.4 LTS, and managed to get it up and running and since this morning, it had been working fine or so I thought......

 

This afternoon, copy backup jobs have been failing - when I have looked at the server, it mentioned IO Buffer error and just now it mentioned Raid control in critical error.

 

 

My Linux usage is basic - so I have no idea who to start diagnosing the issue.

 

I'm wondering if the controller is struggling to cope with the amount if data hitting the discs.....

 

Anyone any troubleshoot ideas for Linux??

 

Cheers

 

Cheers

[ReBoot]

I would start by going back into the RAID management interface you used to setup the array. This may be accessed through the BIOS during boot or there may be a Linux management application. That should give you an idea of the RAID state. From the looks of the errors you may have a failing/failed drive.

 

Its unlikely that Linux will have a application built in to manage the MegaRaid (Avgo) card so you may need to see if the manufacturer provides any support for this. This would be the same for a Windows server. I assume you setup the array using the cards firmware interface accessed through the BIOS. This is probably the best place to start.

 

Something like https://hosting-tutorials.co.uk/tutorials/linux/installing-megaraid-storage-manager-on-debian

Edited October 27, 2021 by ReBoot

[ReBoot]

An afterthought. You may be better off ditching the RAID controller and setting up a ZFS array. This can be managed natively through Linux and I've found it a good solution for my servers.

[mdrabble]

I would start by going back into the RAID management interface you used to setup the array. This may be accessed through the BIOS during boot or there may be a Linux management application. That should give you an idea of the RAID state. From the looks of the errors you may have a failing/failed drive.

 

Its unlikely that Linux will have a application built in to manage the MegaRaid (Avgo) card so you may need to see if the manufacturer provides any support for this. This would be the same for a Windows server. I assume you setup the array using the cards firmware interface accessed through the BIOS. This is probably the best place to start.

 

Something like https://hosting-tutorials.co.uk/tutorials/linux/installing-megaraid-storage-manager-on-debian

 

Sorry for the late reply, was away last week with next to no signal.

 

All disks report fine and no issues being reported by the controller. - worked fine when running as a windows server with all the 23 VMs running from the array.

 

As it is for Veeam Immutable backups, they recommend XFS, so not sure if ZFS is supported - will have to take a look later today.

 

May try removing the array altogether and leave the disks as single disks and see if Ubuntu can handle them better in a software raid.

 

Cheers.

[mdrabble]

Ok, so I changed to Raid Controller from Raid to JBOD mode.

 

Created a Software Raid using ubuntu and all reported fine - until I start copy data to the array.

 

At first I thought it could possibly 2 faulty discs as they appeared to be the same discs each time, but that idea went out the window as tried again this morning and 2 different discs are now showing as removed

 

Cannot see it being the actual card as it worked flawlessly when running with Windows, same goes for the discs and the raid config.

 

Will continue to have a poke around and if come to it, may end up having to re-jog some things around to re-use another server which will be a faff as will need to migrate data across first......