I am also asking myself how to access/mout/"open" the zram-swap manually (it seems to be by default in /var) - for the same reasons as you, btw., but I have not found any info about it yet.
But what I can say is: afaik, the zramfs-thing is compressed; imagine it's encrypted, anyway undisplayable like an (almost?) whole binary file... So I doubt that you will be able to access it as it were some ordinary ext4 partition. That explains why it works to "format" it while your system doesn't get how to mount "this strange thing" (you even can unmount /boot and shred it or "apt purge" via su(do) packages, without which your system won't boot or do anything at all... anyway, you can nuke everything what root can do ("#rm -r"-ing /etc/, /bin, /sbin, the very essential files. BUT only one (and prob. one only) when logged in yet and/or your system has booted / is running. Afterwards -> "booting problems" , ofc if there won't be any issues before. The more essential, the earlier it's loaded into memory/RAM I assume. You don't guess how often I "played around" with something in this context and messed up everything or almost everything (a little off-topic: DO BACKUPS to another drive and/or some webspace/cloud, the latter for the most important files to you; I even automatized this actually nothing but zip/tar/rar and rsync/cp (versioned names or overwrite/replace without any confirmations), KDE since Debian 11 has in it's systemsettings5
a GUI for scheduled and regular backups, I mean some simple gui, nobody needs fancy backup-tools if # crontab -e
and even an one-liner can save your data if booting or else fails badly (always(!!) run crontab -e
with "-e" or your cronta will be resetted!) Although it's strange that it is letting you formatting/changing it even as root...
Anyway, I would advise you to post (the relevant line(s) if not all) of /etc/fstab as well as any other scripts you wrote by yourself or modified with the goal to mount (/tmp or sth more) so people who are more experienced than me may help you if it's possible to find a positive answer to your question.
Also take a look into $ man zramctl
and run $ df -h
and $ mount
(outputs of latter two MAY be of use for a solution). I received now the following:
# sudo swapon --all
# sudo zramctl
NAME ALGORITHM DISKSIZE DATA COMPR TOTAL STREAMS MOUNTPOINT
/dev/zram0 lz4 2G 14.7M 1.7M 2.1M 12 [SWAP]
# sudo swapon --show
NAME TYPE SIZE USED PRIO
/dev/dm-2 partition 980M 0B -2
/dev/zram0 partition 2G 48M 90
Well, you see, mountpoint "[SWAP]"
... can't remember that "conventional" disk-swapping gives you the possibility to access it manually. I guess if some program performs I/O on any (zram-)swap(-partition) the system or even kernel manages what is written and read and where inside any swap-device. Can't find it but I'm pretty sure my zram0 was a file inside /var(subdir...?) at the default size then, back few months ago. dm-2 is the disk-swap which is empty because it's off. Like the manpages tell you, anyway zram#-s are block-devices and disk-swap-devices too, at least I gott that from it's name (dm-2, stands afaik for disk-mapper#2)
Well I rather doubt that we will find a solution we seek for. But if it's possible, I am interested as well. I hope I coulld give you any thoughts, idea or inspiration...