I am wondering how this calculation (8 exbibytes - 1 byte) was computed?
The file size is stored in a signed 64 bit integer, which goes from -2^63 to 2^63-1. Since an exbibyte is 2^60 bytes, 2^63-1 corresponds to 8 exbibytes - 1 byte.
I am also curious if there is a way for me to verify this if I have the filesystem mounted?
Yes, try this:
truncate -s ...
In general, even if the size isn't explicitly stored, two things must be: where the chunks of data are stored, and how big each chunk can be. You take the maximum value for each, and multiply to get the maximum file size. The specifics depend on each filesystem. In XFS the chunks are called "extents" and the number of extents (di_nextents according ...
RAID covers against a drive going bad, and in particular, a drive going bad in a way that very clearly says "I'm broken", either by returning errors to some read request, or just not answering at all.
It doesn't protect against the drive silently corrupting data (they have error correction, of course, but it's still not unheard of), the filesystem ...
In the end, /mnt is a directory like any other. You can create files and directories under it, not just mount(8) stuff there. Also note that if you have, say, /mnt/test and go mount /dev/whatever /mnt, while whatever is mounted there, /mnt/test is totally unacessible. If you reboot, presumably nothing gets mounted there and /mnt/test will be sitting there.
Try looking in your ~/.bash_history (or the appropriate history file, if present).
Otherwise, the mount command with no parameters will tell you whether the directory is indeed mounted (also check /etc/fstab just in case).
You might also try looking (using find) in the whole system for files changed or created more or less at the same time as your /mnt/test.
You can't find out what mounted it, but you can see if it's still mounted by looking at the output of mount or even just attempting to remove the mountpoint, rmdir /mnt/test (success means nothing was mounted there).
To unmount a filesystem mounted there, umount /mnt/test and then rmdir /mnt/test
an error in the upper ext4 filesystem would affect both drives simultaneously because both drives are readed and written at the same time. Is there any means to prevent this?
None that I'm aware of.
By flexibility I mean, what happens if one of the mirrored drives breaks up and I'm not able to afford another drive? can I maintain a RAID1 system with only ...
Yes, bind mount is working.
I just didnt understand at the first.
But after looking into /proc/self/mountinfo, i got the better understanding.
shared:X Optional field is the catch and Mount ID, Parent ID also gave the clue.
47 14 0:30 /hostname /etc/hostname rw,relatime shared:21 - ...
You can watch current block device I/O traffic with nmon, as in:
NMON=ld nmon -s1
(this preselects display of load graph and disk graph with 1 second refresh time.. this can also be set by starting nmon without parameters, than pressing l, d, - to toggle graphs and reduce refresh time from default 2s .. + of course would increase refresh delay.)