-
タイトル
Unreadable (pending) sectors -
説明
Warning!This knowledge article applies to the following unsupported hardware appliances:
- SCB N1000(d)
- SCB N5000
- SCB N10000
- SSB N1000(d)
- SSB N5000
- SSB N10000
There are pending sector alerts, which comes from the SSB/SCB system.
-
原因
Usually these are just warnings that there is one or more sectors on the disk that the HDD failed to read.
Until it can read properly, the disk will flag it as a pending (weak) sector. -
対策
Only a write to the specified sector can decide whether it is a bad or a good sector.
We suggest to do a RAID array re-check to help the disk deciding its state.n1000, n1000d
As our n1000 and n1000d contains software RAID with two disks, the recheck can be initiated from the boot shell using a script.
1. Log in to the boot shell on the given node
2. Run the RAID check
/usr/share/mdadm/checkarray -a
3. To see the check's state (exit with Ctrl + C)
watch -n1 cat /proc/mdstat
n5000, n10000
These appliances contain hardware RAID controllers, so you have to ask the controller to do the recheck.
This part of the guide checks the logical drive's consistency.1. Log in to the boot shell on the given node
2. Run the RAID check
arcconf TASK START 1 LOGICALDRIVE 0 VERIFY_FIX
3. Monitor the RAID controller's state
arcconf getstatus 1
Additionally checking each disk separately for any media errors is also advised and only differs at Step #2:
arcconf TASK START 1 DEVICE 0 DISK_ID VERIFY_FIX
The DISK_ID should be between 8 and 19, but please ask Support to help finding out the exact numbers.