Failing drive(s) notification after 9.3 upgrade

[slyph]

Hello,

My ignorance has been bliss and I hadn't administered my NAS in...too long. I recently upgraded to 9.3 and now I'm being notified that my drives are bad (among other things). Again, I didn't follow the recommendations for regular long SMART tests, so below links are the results from their first ever run last night. I'm fairly sure that I need to replace a drive or two. Could I get some feedback on which drives I should replace?

zpool status
smartctl -x /dev/ada0
smartctl -x /dev/ada1
smartctl -x /dev/ada2
smartctl -x /dev/da0
smartctl -x /dev/da1
smartctl -x /dev/da2
smartctl -x /dev/da3
smartctl -x /dev/da4
smartctl -x /dev/da5
smartctl -x /dev/da6

//Edit
Hardware Configuration:
FreeNAS-9.3-STABLE-201505130355
GIGABYTE GA-Z77X-UD5H LGA 1155 Intel Z77
Intel Core i5-3450S Ivy Bridge 2.8GHz (3.5GHz Turbo) LGA 1155 65W Quad-Core
8x SAMSUNG 4GB 240-Pin DDR3 SDRAM DDR3 1600 (PC3 12800) Desktop Memory Model MV-3V4G3/US
7x Seagate Barracuda ST3000DM001 3TB 7200 RPM 64MB Cache SATA 6.0Gb/s 3.5"
2x Hitachi Deskstar 5K3000-HDS5C3030ALA630 3TB 7200 RPM
M1015 Controller - on firmware v14
NORCO RPC-4020 4U Rackmount Server Chassis w/ 20 Hot-swappable SATA/SAS Drive Bays

I really appreciate the feedback!

 

[rogerh]

Happened to be passing: those qualified to comment will be helped by info about your hardware and software as per forum rules.

 

[danb35]

• All of your drives are too hot. Move your server to a cooler environment and/or increase airflow over the drives.
• You need to upgrade the firmware on your HBA to version P16. Do not use a newer version--the current firmware from LSI is P20, but the firmware version needs to match the driver version, and the driver version in FreeNAS is P16.
• ada0, ada1, and da5 are all showing bad sectors and should probably be replaced. Of these, da5 is in the worst shape so should be replaced first. da5 and ada0 are both failing long SMART tests, while ada1 isn't. You might be able to get away with holding on to ada1 for a little while longer, but I'd be planning to replace it as well.
• It's probably going to be difficult for you to change this (since it would involve backing up all your data, destroying your pool, recreating it, and restoring the data), but your pool configuration really isn't ideal. You'd be better off with a single RAIDZ2 vdev (which would also increase your capacity).

The manual has click-by-click instructions for replacing a failed disk. Follow them. Do not, under any circumstances, use the volume manager--that has nothing to do with replacing disks and will result in losing redundancy on your pool.