You are not logged in.
Pages: 1
Hi !
I just installed Archlinux on a new computer and I decided to try Btrfs.
My OS drive is
/dev/nvme0n1
, partitionned this way:
BIOS Boot Partition
EFI System Partition
Linux Btrfs
The Btrfs partition contains two subvolumes:
/@root
/@swap
Everything is mounted this way:
# Static information about the filesystems.
# See fstab(5) for details.
# <file system> <dir> <type> <options> <dump> <pass>
# /dev/nvme0n1p3
UUID=71491f6f-22f2-4da0-82f5-9d4a5871d5f9 / btrfs rw,relatime,ssd,space_cache,subvolid=256,subvol=/@root,subvol=@root 0 0
# /dev/nvme0n1p2 LABEL=EFI
UUID=8409-787B /efi vfat rw,relatime,fmask=0022,dmask=0022,codepage=437,iocharset=ascii,shortname=mixed,utf8,errors=remount-ro 0 2
# /dev/nvme0n1p3
UUID=71491f6f-22f2-4da0-82f5-9d4a5871d5f9 /swap btrfs rw,relatime,ssd,space_cache,subvolid=257,subvol=/@swap,subvol=@swap 0 0
/swap/swapfile none swap defaults 0 0
Everything is working well ; system is booting without errors and I'm able to auto mount when needed using
mount -a
Then I wanted to try RAID 1 capability of Btrfs, using two spare drives.
I chose a partitionless setup:
mkfs.btrfs -d raid1 -m raid1 -L "NVMe Data" /dev/nvme1n1 /dev/nvme2n1
Then I created two subvolumes:
/@postgres
/@clickhouse
And I appended the following to my fstab:
# @postgres
UUID=d91798c7-a5fc-401a-a34c-59cfc371611a /var/lib/postgres btrfs rw,relatime,ssd,space_cache,subvolid=258,subvol=/@postgres 0 0
# @clickhouse
UUID=d91798c7-a5fc-401a-a34c-59cfc371611a /var/lib/clickhouse btrfs rw,relatime,ssd,space_cache,subvolid=259,subvol=/@clickhouse 0 0
After a restart, the subvolumes were properly mounted. However, each
mount -a
while the subvolumes are already mounted will lead to an error:
mount: /var/lib/postgres: /dev/nvme2n1 already mounted or mount point busy.
mount: /var/lib/clickhouse: /dev/nvme2n1 already mounted or mount point busy.
In my understanding of mount, it should not raise an error if the subvolume are already mounted. Am I missing something ?
I'm using 5.11.15-arch1-2 and I disabled my postgres & clickhouse servers to prevent any files usage from these path.
Thanks for your help.
Offline
After further testing, it looks like mount doesn't like the usage of UUID for RAID1 Btrfs ...
Replacing UUID=d91798c7-a5fc-401a-a34c-59cfc371611a by either /dev/nvme1n1 or /dev/nvme2n1 does the tricks but I would prefer to use UUID...
Offline
I don't know about your actual problem. It should work with UUID.
Something else:
You should remove all of those "subvolid=..." mount options you are using. Use only "subvol=...".
EDIT:
I think I understand what's going on now: you are right now confused about what subvolumes are and how they work. You seem to have created your new postgres and clickhouse subvolumes inside your "UUID=71491f6f-22f2-4da0-82f5-9d4a5871d5f9" filesystem. You need to mount your "UUID=d91798c7-a5fc-401a-a34c-59cfc371611a" filesystem and then create the postgres etc. subvolumes inside that filesystem.
Also, I guess you'll want to name your subvolume "@var@lib@postgres" because you'll mount it to "/var/lib/postgres"? I don't know how that naming scheme works that uses "@" characters. I thought the "@" is supposed to mean "/" so a "@postgres" name means you mount something to "/postgres"? The "@" characters are somehow important for software like timeshift and you use timeshift? Maybe you can find something in the timeshift documentation about what they want you to do with the "@" characters. The only thing I could find is that timeshift cares about subvolumes "@" and "@home" for "/" and "/home". Maybe it doesn't care about any other subvolume so you don't really need those "@" characters for your postgres subvolume? I would then just call it "postgres" without that weird "@".
Last edited by Ropid (2021-04-21 01:38:39)
Offline
You seem to have created your new postgres and clickhouse subvolumes inside your "UUID=71491f6f-22f2-4da0-82f5-9d4a5871d5f9" filesystem. You need to mount your "UUID=d91798c7-a5fc-401a-a34c-59cfc371611a" filesystem and then create the postgres etc. subvolumes inside that filesystem.
Yes, I agree.
Offline
Pages: 1