Delay fsck routine disc checks until after OS has booted
So say you have a file server. It's OS resides on /dev/sda. However, all of the data resides on MDADM arrays, /dev/md###. The number of different arrays is, say for arguments sake, five.
Now, we know that fsck will run a routine check on partitions once every 120 (?) days or every ### (?) mounts. The numbers aren't important in this scenario.
If a system administrator needs to reboot the server after, say, 200 days, clearly, all disks will want a routine check. So when the system reboots (after maintenance or what have you), instead of the system being down for the time it takes to reboot, now it's down for the time it takes to reboot PLUS the time it takes to scan all the partitions/disks/arrays. This can be an incredibly long downtime especially if there are huge partitions/arrays and lots of data to check.
How can we configure fsck to DELAY it's scan until after the system and services come online?
The best example here is:
* Say that all disks were recently scanned, except one, which has had it's routine scheduled check duration pass by, so next time we reboot, the system will scan this array for errors.
* When we reboot, wouldn't it be ideal to have the entire system come up (SSH, web services, the rest of the file shares, etc) and only have the array scheduled for scanning be temporarily down? indeed ;-0