You are not logged in.
Hi.
I'm reinstalling Arch after finally receiving a replacement hard-drive. I'm considering holding off creating a Btrfs for / until the "snappy" compression algorithm becomes available. Can anyone tell me if it's possible to do now?
My last install used a lzo compressed btrfs root.
Offline
The official package is Not yet updated with that. It needs kernel support too.
http://www.phoronix.com/scan.php?page=n … px=MTA0MjQ
Offline
They should focus on implementing an fsck and a recovery tool first. That's the biggest drawback from using btrfs right now, not disk compression. One power outage and you'll have to reinstall Arch!
I have made a personal commitment not to reply in topics that start with a lowercase letter. Proper grammar and punctuation is a sign of respect, and if you do not show any, you will NOT receive any help (at least not from me).
Offline
@DS - I thought that the fsck support was merged in mainline 3.3? Looks like [core]/btrfs-progs are dated "0.19.20120110-2" and there is nothing in [testing] that is newer.
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
Oracle declared btrfs to be stable few days ago.
http://groups.google.com/group/snappy-c … 7234c161cd --> look at the snappy release date.
Arch package says its unstable and it is the current version.
Offline
They should focus on implementing an fsck and a recovery tool first. That's the biggest drawback from using btrfs right now, not disk compression. One power outage and you'll have to reinstall Arch!
Am I missing something? I seem to have a btrfsfsck... haven't needed to use it yet...
# whereis btrfsck
btrfsck: /usr/bin/btrfsck /usr/share/man/man8/btrfsck.8.gz
for some reason it's fsck isn't picked up when I do a mkinitcpio...
# mkinitcpio -p linux
==> Building image from preset: 'default'
-> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux.img
==> Starting build: 3.2.12-1-ARCH
-> Parsing hook: [base]
-> Parsing hook: [udev]
-> Parsing hook: [autodetect]
-> Parsing hook: [pata]
-> Parsing hook: [scsi]
-> Parsing hook: [sata]
-> Parsing hook: [filesystems]
-> Parsing hook: [usbinput]
-> Parsing hook: [btrfs]
-> Parsing hook: [btrfs_advanced]
-> Parsing hook: [fsck]
==> ERROR: file not found: `fsck.btrfs'
==> WARNING: No fsck helpers found. fsck will not be run on boot.
==> Generating module dependencies
==> Creating gzip initcpio image: /boot/initramfs-linux.img
==> Image generation successful
==> Building image from preset: 'fallback'
-> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux-fallback.img -S autodetect
==> Starting build: 3.2.12-1-ARCH
-> Parsing hook: [base]
-> Parsing hook: [udev]
-> Parsing hook: [pata]
-> Parsing hook: [scsi]
-> Parsing hook: [sata]
-> Parsing hook: [filesystems]
-> Parsing hook: [usbinput]
-> Parsing hook: [btrfs]
-> Parsing hook: [btrfs_advanced]
-> Parsing hook: [fsck]
==> Generating module dependencies
==> Creating gzip initcpio image: /boot/initramfs-linux-fallback.img
==> Image generation successful
Offline
Oracle declared btrfs to be stable few days ago.
Oracle says so many things...
Got Leenucks? :: Arch: Power in simplicity :: Get Counted! Registered Linux User #392717 :: Blog thingy
Offline
==> ERROR: file not found: `fsck.btrfs'
I had this error too. But didnt know how to deal with that, and installed the OS in ext4.
Offline
They should focus on implementing an fsck and a recovery tool first. That's the biggest drawback from using btrfs right now, not disk compression.
We have both in btrfs already: the repairing version of btrfsck is in Chris' git repo and is included with btrfs-progs on Oracle Linux 5 and 6. Likewise, we now have both a read-only recovery tool that can copy data off a damaged btrfs volume, as well as a recovery mount option (mount -o recovery) that will back-walk the btrfs b-trees to find a stable, consistent filesystem.
Offline
I'm reinstalling Arch after finally receiving a replacement hard-drive. I'm considering holding off creating a Btrfs for / until the "snappy" compression algorithm becomes available.
Is there a particular reason you're waiting for Snappy? Keep in mind that you can actually change compression algorithms on-the-fly, i.e. start with LZO now and switch to Snappy compression later. Obviously older files that are compressed will remain LZO, but new files and modified files will be recompressed with Snappy.
Offline
I'm running an old PC with no hope of upgrade for now with the bottleneck being a Pentium 4 cpu.
This was based off a completely subjective feel, and while I didn't actually do any benchmarks, I thought the following.
- Uncompressed Btrfs performed better on desktop tasks than Ext4 did.
- Btrfs with lzo compression created a noticible lag time when loading apps.
I was hopingthe snappy compression would create a nice compromise while trying to save as much space as possible.
Offline
Is btrfs compression even beneficial if you have an ssd that already does compression (like intel ssd's)?
Offline
DSpider wrote:They should focus on implementing an fsck and a recovery tool first. That's the biggest drawback from using btrfs right now, not disk compression.
We have both in btrfs already: the repairing version of btrfsck is in Chris' git repo and is included with btrfs-progs on Oracle Linux 5 and 6. Likewise, we now have both a read-only recovery tool that can copy data off a damaged btrfs volume, as well as a recovery mount option (mount -o recovery) that will back-walk the btrfs b-trees to find a stable, consistent filesystem.
Avi - I can't seem to locate the repo to which you refer. I did find the following, but this hasn't seen an update in 6 months. Can you post the url? Would love to get my hands on the fsck util
CPU-optimized Linux-ck packages @ Repo-ck • AUR packages • Zsh and other configs
Offline
Avi - I can't seem to locate the repo to which you refer.
https://git.kernel.org/?p=linux/kernel/ … ads/master
This is actually the version of btrfs-progs we've released as part of the UEK2 production release on Oracle Linux. Chris merged it across this morning/evening depending on your timezone.
Last edited by Avi Miller (2012-03-28 20:17:35)
Offline
I have an intel 520 (cherryville) that I was thinking about installing arch with btrfs compress=lzo, would there be any benefit to this or would I be defeating the purpose of the on ssd compression. I tried the install a few times with and without, it took a long time to open an encrypted document and some vm's seemed laggy. It was less so without, but I'm not sure that was a misconfiguration or not. I even tried a gentoo install with out it and it was huge on the disk, I might try it with compression to see what difference it would make.
Besides my little experience there doesn't seem to be a lot of definitive info about compression with btrfs on an ssd that does compression, whether it helps or not.
Offline
Using compression for an SSD means less data to be written and thus, less wear on the cells. But compressing an already compressed file could result in a slightly larger file (not to mention a higher CPU involvement in the process) - which could very well be pointless.
But I think that this discussion is outside the scope of this thread.
If your SSD comes with some form of hardware compression, maybe you can disable it with a firmware tool, setting, reflash, something and let the OS compress them with a better algorithm/method.
I have made a personal commitment not to reply in topics that start with a lowercase letter. Proper grammar and punctuation is a sign of respect, and if you do not show any, you will NOT receive any help (at least not from me).
Offline