>
> Ok, so I will try to force assemble my array.
> However, smartctl -x reports:
> 5 Reallocated_Sector_Ct PO--CK 100 100 036 - 38
> So my disk is probably not as clean as expected, isn't it?
>
I would not worry about the reallocated sectors.
So long as the disk either reallocates the sectors or rewrites and
fixes things are good, so long as there are not other issues.
I did have one disk that never seemed to reallocate/fix all of them
and would every so often give a 7 second pause.
I have also had a couple of disks that ran out of space in the
reallocation area. From my experience the number of reallocations a
disk can do depends on the model, so if the number starts rapidly
rising then you should probably plan for a new disk.
I also run this in cron daily, to collect smart reports so I can check patterns.
stamp=`date +%Y%m%d-%H`
for disk in a b c d e f g h i j k l m n o p ; do
smartctl --all /dev/sd${disk} > /var/log/smartctl/tmp.out
serial=`grep Serial /var/log/smartctl/tmp.out | awk '{print $3}'`
mkdir -p /var/log/smartctl/${serial}
mv /var/log/smartctl/tmp.out
/var/log/smartctl/${serial}/${serial}.${stamp}.sd${disk}.out
if [ "${disk}" == "a" ] ; then
smartctl -l ssd /dev/sd${disk} >>
/var/log/smartctl/${serial}/${serial}.${stamp}.sd${disk}.out
fi
if [ $? -eq 2 ] ; then
rm -f /var/log/smartctl/sd${disk}.${stamp}.out
fi
done
And I have also learned that once a disk starts giving you trouble,
get it clean of these 2:
197 Current_Pending_Sector 0x0012 100 089 000 Old_age
Always - 0
198 Offline_Uncorrectable 0x0010 100 089 000 Old_age
Offline - 0
And run this daily:
smartctl -t long /dev/sdXX
Doing the long test often enough seems to let the disk find and
rewrite the sectors before they have a serious issue, doing it weekly
was not enough to keep it clean. And I only run this daily on the
disks that have had problems. Weekly I run a long test on all
disks, and I do weekly scrubs, and will force a scrub if I see issues.
Good luck.
