We accidentally removed a SAS cable from an active AVAGO MegaRAID controller during a troubleshooting session. After we reconnected the SAS cable, all disks in the RAID6 disk array are reported as Frn-Bad and the virtual drive is gone.
We could now change all drives to "unconfigured good" and then create a new virtual drive, but there was still some (non-essential) data on the JBOD. Is it possible to recreate the disk configuration without losing data, or is the data irrecoverably lost?
TL;DR: If you accidentally removed a SAS cable from your server then install storcli and run
storecli /call show
note the correct controller and enclosure and the nrun
#!/bin/bash
# Use your values from above
CONTROLLER=
ENCLOSURE=
storcli /c$CONTROLLER /e$ENCLOSURE /sall set good
storcli /c$CONTROLLER /fall show
echo -n "Import all foreign drives? [y/n]"
read yno
case $yno in
[yY] | [yY][Ee][Ss] )
storcli /c$CONTROLLER /fall import
;;
*) echo "Skipping import"
;;
esac
When you disconnected the SAS cable, the controller noticed that all drives where gone and therefore potentially in a bad state. That's why you get Bad. Also, the drives were potentially new and not yet included in your RAID setup, that's why you get Frn and as a result Frn-Bad.
In order to fix this, you first need to mark all drives as UGood (unknown, but good). The simplest way is to use storcli (download from LSI). First look up all drives with
storcli /call show
The /call will look for all controllers, but you might as well use /cX for a specific one. All Frn-Bad drives will reside in a single enclosure, e.g.
8:1 | drive information .... | Frn-Bad | ...
8:13 | drive information .... | Frn-Bad | ...
8:14 | drive information .... | Frn-Bad | ...
8:23 | drive information .... | Frn-Bad | ...
The common number in front of the colon is the enclosure. Call storcli again and use the enclosure number to verify that you really have the right drives at hand:
storcli /call /e8 /sall show
Now set all those drives to good
storcli /call /e8 /sall set good
This will mark the drives as unknown and good. The MEGARaid controller saves the configuration on the hard drives. You can show it with
storcli /c0 /fall show
and then import it with
storcli /c0 /fall import
It should take less than a minute to import the drives.
It could be useful to perform this check:
https://www.servethehome.com/fixing-drive-labeled-foreign-lsi-avago-sas-controller/
Unfortunately I hope the configuration was not fully removed. If you able to scan the foreign configuration, you should still be able to enable the Raid Array.
But I cannot guarantee this as you have already created a new Virtual drive with those disks.
