Back to the main page
Replacing hard disk on StorEdge 6120
Basically you'd like to see ' Blue - safe to remove' LED on failed HDD before remove it, but you may not be lucky always.
The failed HDD usually has 'Amber - service action required' LED and maybe ' Green ' one.
So you can try getting Blue LED with the command disable hdd where hhd is something like u1d03.
But if HDD is already failed, then array cannot communicate with it and no luck to see Blue LED, so you have no choice then to continue with replacement.
Another command can be tried: vol disable u1d03 to_standby , when data is first reconstructed (or copied) onto the hot spare (standby) and then u1d03 is disabled.
Actually this one is good if you replace good drive, say replace smaller drive with bigger one. In this case, after data is copied to spare drive, you should see Blue LED on disk to be replaced.
- Establish console connection to 6120 - hopefully you are using Cyclades console servers, if not find more about it :)
- Push down on latch release to open hdd latch handle
- Pull out the HDD about 2-3cm and wait for 20-30 sec for hdd to spin down (okay I know you don't have time to wait even 30sec :)
-
- Remove hdd from the array (6120)
- Wait for say 1 min so array can figure out new state
- Install new hdd
- If needed, run enable u1d03
- Use next commands to verify new hdd
- fru stat
- fru list
- vol stat
- vol list
- Also check proc list to see % of reconstruction of RAID or volume.
Basically once hdd from RAID fails, hot spare replaces it and reconstruction begins.
You have to wait to be finished, before new reconstruction will begin with new hdd.
Post csript
One day I wanted to replace all disks in the 6120, to upgrade all of them to 300G.
So I had to do one by one. Fist I replaced spare one, disk #14.
With data disks, I was doing like below, say example for u1d03:
Check drive:
All drives mounted.
# vol stat
vol1: mounted
u1d01: mounted
u1d02: mounted
u1d03: mounted
u1d04: mounted
u1d05: mounted
u1d06: mounted
u1d07: mounted
u1d08: mounted
u1d09: mounted
u1d10: mounted
u1d11: mounted
u1d12: mounted
u1d13: mounted
Standby: u1d14: mounted
|
# fru list u1d03
ID TYPE VENDOR MODEL REVISION SERIAL
------ ----------------- ----------- ----------- ------------- --------
u1d03 disk drive SEAGATE ST336753FSUN 0349 3HX1EZPM
|
# fru stat u1d03
DISK STATUS STATE ROLE PORT1 PORT2 TEMP VOLUME
------ ------- ---------- ---------- --------- --------- ---- ------
u1d03 ready enabled data disk ready ready 31 vol1
|
No volume reconstruction in progress.
# proc list
VOLUME/DISK CMD_REF PERCENT TIME COMMAND
vol1 37952 2 12:48 vol verify
|
I will replace this drive (36G) with bigger one (300G).
This command will copy (reconstruct) data to standby (spare) disk and then (after copy is done) disable disk to be replaced.
# vol disable u1d03 to_standby
|
Say, u1d14 is standby disk. The u1d03 and u1d14 will have Green LED fast blinking.
Other drives will have Green LED constantly on, no blinking.
I wait 15-20 min for copy to be finished.
Now I see that disk is substituted with spare one.
# vol stat
vol1: mounted
u1d01: mounted
u1d02: mounted
u1d03: substituted by u1d14
u1d04: mounted
u1d05: mounted
u1d06: mounted
u1d07: mounted
u1d08: mounted
u1d09: mounted
u1d10: mounted
u1d11: mounted
u1d12: mounted
u1d13: mounted
Standby: u1d14: mounted
|
And disk is not available any more.
# fru list u1d03
ID TYPE VENDOR MODEL REVISION SERIAL
------ ----------------- ----------- ----------- ------------- --------
u1d03 disk drive - - - -
|
# fru stat u1d03
DISK STATUS STATE ROLE PORT1 PORT2 TEMP VOLUME
------ ------- ---------- ---------- --------- --------- ---- ------
u1d03 fault substituted
|
Disk u1d03 has Blue LED.
I am taking u1d03 out, you check correct way of doing this!
I input bigger disk (300G)
Now I see that data is copied back (reconstructed) from standby disk to bigger one.
Again, heavy blinking of Green LED only on u1d03 and u1d14 (standby disk).
# proc list
VOLUME/DISK CMD_REF PERCENT TIME COMMAND
vol1 52290 16 1:32 vol recon
|
Storage can see bigger disk with state 'substituted'.
# list u1d03
ID TYPE VENDOR MODEL REVISION SERIAL
------ ----------------- ----------- ----------- ------------- --------
u1d03 disk drive FUJITSU MAT3300FC 0108 P56002D7
|
# fru stat u1d03
DISK STATUS STATE ROLE PORT1 PORT2 TEMP VOLUME
------ ------- ---------- ---------- --------- --------- ---- ------
u1d03 ready substituted data disk ready ready 21 vol1
|
Wait again for reconstruction to finish.
After reconstruction is finished, disk is mounted again.
# vol stat
vol1: mounted
u1d01: mounted
u1d02: mounted
u1d03: mounted
u1d04: mounted
u1d05: mounted
u1d06: mounted
u1d07: mounted
u1d08: mounted
u1d09: mounted
u1d10: mounted
u1d11: mounted
u1d12: mounted
u1d13: mounted
Standby: u1d14: mounted
|
Now disk status is "enabled"
# fru stat u1d03
DISK STATUS STATE ROLE PORT1 PORT2 TEMP VOLUME
------ ------- ---------- ---------- --------- --------- ---- ------
u1d03 ready enabled data disk ready ready 26 vol1
|
Now I can repeat all these steps for each disk until the storage has all disks 300G. True, it will take some time.
Back to the main page