You are not logged in.
Pages: 1
When insatlling arch on a btrfs formatted partition, I found the boot to be very slow compared to ext4, I found out why:
https://bugs.launchpad.net/ubuntu/+sour … omments/27
any btrfs partition should have a '0' for 'pass' in the fstab
currently the genfstab script gives it a '1'
So I think that genfstab should have a '0' for pass when a partition is btrfs?
Should I create a bug report? (just want to confirm)
bitcoin: 1G62YGRFkMDwhGr5T5YGovfsxLx44eZo7U
Offline
Would seem like a valid bug report to me, but from my understanding the boot shouldn't actually be slowed down because there isn't an fsck.btrfs binary to execute unless you symlinked it yourself (which is a bad idea). You may however get an error message on boot about it.
Offline
Would seem like a valid bug report to me, but from my understanding the boot shouldn't actually be slowed down because there isn't an fsck.btrfs binary to execute unless you symlinked it yourself (which is a bad idea). You may however get an error message on boot about it.
I got an error when running:
mkinitcpio -p linux
but no errors in boot or in journal from what I can remeber.
I didnt do any symlinking, but still had a slow boot, unfortunately I formatted with ext4 now.
bitcoin: 1G62YGRFkMDwhGr5T5YGovfsxLx44eZo7U
Offline
Change fstab and see if the boot gets faster.
Edit:
unfortunately I formatted with ext4 now.
I see.
Last edited by karol (2013-05-26 11:00:00)
Offline
65kid wrote:Would seem like a valid bug report to me, but from my understanding the boot shouldn't actually be slowed down because there isn't an fsck.btrfs binary to execute unless you symlinked it yourself (which is a bad idea). You may however get an error message on boot about it.
I got an error when running:
mkinitcpio -p linux
but no errors in boot or in journal from what I can remeber.
I didnt do any symlinking, but still had a slow boot, unfortunately I formatted with ext4 now.
if you have btrfs as root you should simply remove the fsck hook from your /etc/mkinitcpio.conf (and set passno to 0 of course).
Offline
I will test again with btrfs when I have time soon, I quite like btrfs
bitcoin: 1G62YGRFkMDwhGr5T5YGovfsxLx44eZo7U
Offline
Much faster without the fsck hook and making 'pass' '0' in fstab. Ill file a bug report and make en entry in the btrfs wiki
bitcoin: 1G62YGRFkMDwhGr5T5YGovfsxLx44eZo7U
Offline
Much faster without the fsck hook and making 'pass' '0' in fstab. Ill file a bug report and make en entry in the btrfs wiki
The only way this makes sense is if you've created some bogus symlink called fsck.btrfs pointing to btrfsck. Otherwise, it should simply fail and move on.
Offline
jrussell wrote:Much faster without the fsck hook and making 'pass' '0' in fstab. Ill file a bug report and make en entry in the btrfs wiki
The only way this makes sense is if you've created some bogus symlink called fsck.btrfs pointing to btrfsck. Otherwise, it should simply fail and move on.
for now it seems quicker with pass '0', I just wish I had not formatted but I will test a bit later.
bitcoin: 1G62YGRFkMDwhGr5T5YGovfsxLx44eZo7U
Offline
cat /etc/fstab
#
# /etc/fstab: static file system information
#
# <file system> <dir> <type> <options> <dump> <pass>
# /dev/sda3
UUID=db2b039e-e606-4981-94f7-b2a68139c255 / btrfs rw,relatime,compress=lzo,space_cache 0 1
# /dev/sda2
UUID=bc1266e5-6f2b-4d23-ad73-47bf5608b5b1 none swap defaults 0 0
cat /etc/mkinitcpio.conf
HOOKS="base udev autodetect modconf block filesystems keyboard btrfs resume"
I don't think it matters. I know there is no need of 1 but I just didn't bother. It works fine on my system.
Offline
cat /etc/fstab # # /etc/fstab: static file system information # # <file system> <dir> <type> <options> <dump> <pass> # /dev/sda3 UUID=db2b039e-e606-4981-94f7-b2a68139c255 / btrfs rw,relatime,compress=lzo,space_cache 0 1 # /dev/sda2 UUID=bc1266e5-6f2b-4d23-ad73-47bf5608b5b1 none swap defaults 0 0
cat /etc/mkinitcpio.conf HOOKS="base udev autodetect modconf block filesystems keyboard btrfs resume"
I don't think it matters. I know there is no need of 1 but I just didn't bother. It works fine on my system.
Im going to fiddle a bit more when I have time, Id like to see what the problem/cause was.
I agree that it shouldn't make a difference, and that it will just fail and pass. But I noticed that systemd-analyze gave around 6s for entire boot, but there was lots of hdd chatter after the login prompt (in multi-user with nothing else enabled) for about another 3-4s with btrfs comparted to nearly zero hdd noise after the promt, also with around 6s on ext4.
Perhaps I did symlink
Last edited by jrussell (2013-05-29 15:47:50)
bitcoin: 1G62YGRFkMDwhGr5T5YGovfsxLx44eZo7U
Offline
Offline
cat /etc/mkinitcpio.conf HOOKS="base udev autodetect modconf block filesystems keyboard btrfs resume"
I don't think it matters. I know there is no need of 1 but I just didn't bother. It works fine on my system.
Sorry about this post not relating to the actual topic of the thread. But I just wanted to point this out to donniezazen.
==> Help for hook 'btrfs':
This hook provides support for multi-device btrfs volumes. This hook
is only needed for initramfs images which do not use udev.==> This hook has runtime scripts:
-> pre-mount hook
Apparently, the btrfs hook is not hurting anything. But it is definitely not needed.
Offline
jrussell wrote:Perhaps I did symlink
I'm quite sure you did. Regardless, this is fixed in the arch-install-scripts package (v11) I pushed yesterday, so it'll be on the June ISO.
Cool, thanks.
It was stupid of me to format like that so quickly, oh well, keep learning.
bitcoin: 1G62YGRFkMDwhGr5T5YGovfsxLx44eZo7U
Offline
Pages: 1