Mount multiple disks in one location without losing all of the data if one disk breaks?


Setting up the network shares has led me to create mount points such as “/Shared/SeriesI” and “/Shared/SeriesII”, now these show up over the (samba) network as SeriesI and SeriesII as well.
Can I merge these into one big folder? Both are 1TB drives, can I turn them into a 2TB folder/mount point somehow where not all 2TB of data disappears if one of the disks breaks?

In the land of Oz - probably. How do you imagine this? Where data should come from if disk it is stored on fails?

You can have 1TB of redundant storage which will survive single drive failure.

If they have no data on them they could be reconfigured as RAID 0 (striped and 2TB) or RAID 1 (mirrored and 1TB) but you would still need a 1 or 2TB backup disk (mirror is not backup, striped would loose the data) and a 1TB for a replacement on failure.

It also depends on the motherboard controller, since if you have the same mirror on the one controller and this fails, then you would loose the mirror until it could be fixed…

I am not sure if I understand fully what you mean.

You can not mount two file systems on the same mount point. That is you can, but you will only see the data of the last mount.

When you have e.g. a directory /mnt/testmnt with files in it and you then use it as a mount point mounting the file system in /dev/sdf1 on it, you will then see in /mnt/testmnt all that is in the file system in /dev/sdf1. You will not see anymore the original files that are in /mnt/testmnt (but they will reappear when you unmount /dev/sdf1).

Same when you then mount the file system in /dev/sdq5 on /mnt/testmnt. You will then see all that is in the file system of /dev/sdq5 and not anymore those of the file system in /dev/sdf1.

I have no idea if this sounds very logical to you and thus my explanation is superfluous, but I got the idea you are confusing things here. Sorry if I am wrong.

Actually I probably misunderstood the question. You can mount them as …/Series/SeriesI and …/Series/SeriesII and share parent catalog …/Series. This way clients won’t need to change there drive mappings. This allows you to add any number of nested mount points later.

If you literally want to have flat directory without nested folders - using btrfs with raid1 metadata may work (except there is known issue with mounting degraded filesystem).

Assuming you are setting this up and not committed or have reliable backups. You could use LVM containers linked so they appear to be one large 2 TB container with one large file system and then set directories to separate seriesI from seriesII. You can’t just merge two random partitions without preplaning though

Thank you for all the replies, I probably could’ve done a better job explaining this.

I’m aware of RAID 0/1/0+1. Raid 0’s benefit of seeing things as one storage point is what I’m after. However I don’t want it so badly that I want its disadvantage where all data disappears if a single disk fails.

In the land of Oz - probably. How do you imagine this? Where data should come from if disk it is stored on fails?

I’d just want half the data to survive ;). It’s not vital data so it’s not worth the costs of an extra disk for me to store things redundantly, it is however a bit of a hassle to get back so I’d like at least some to remain.

To be honest it doesn’t bother me at all, it’s just my girlfriend who likes things super tidy and doesn’t like seeing 2 network shares with the same purpose (and more clicks as your later nested directories suggestion would result in are out of the question as well).

@gogalthorp I’ll look into that, thank you. At the moment I still have the space to backup things to another disk to organise/reformat things.