Tag Info

Hot answers tagged

5

It's been a while since I played with zfs, but you should be able to use zfs list -t snapshot to find your available snapshots and access the files under a special .zfs directory under your zfs mountpoint. [~]# zfs list -t snapshot NAME USED AVAIL REFER MOUNTPOINT mypool 1.49G 527M 528M /mnt/zfspool ...


3

You can add devices to a pool after it has been created, however not really in the way you seem to envision. With ZFS, the only redundant configuration that you can add devices to is the mirror. It is currently not possible to grow a raidzN vdev with additional devices after it has been created. Adding devices to a mirror increases the redundancy but not ...


2

1: In your configuration, btrfs should detect data corruption but won't correct it as there is no data redundancy as far as btrfs is concerned. 2: RAID6 doesn't detect nor reliably protect against bit corruption. It only protects against disk failure and in some cases might detect unused still bad blocks. 3: If btrfs RAID6 is not ready, you might consider ...


2

This is a bug in the new ZFS Volume Manager in FreeNAS version 9.1.x. It looks like it will be fixed in (hopefully) the next release: https://bugs.freenas.org/issues/3274#note-3


2

This is the best answer, with theory of probabilities too: http://evadman.blogspot.com/2010/08/raid-array-failure-probabilities.html?showComment=1337533818123#c7465506102422346169


2

According to Oracle's zfs documentation, A snapshot is a read-only copy of a file system or volume. Snapshots can be created almost instantly, and they initially consume no additional disk space within the pool. However, as data within the active dataset changes, the snapshot consumes disk space by continuing to reference the old data, thus ...


1

UFS is less resource intensive, ZFS is more resource intensive, usually needing a bit more RAM, so it's not a great fit for VMs. Beyond that, at least for me, ZFS is preferred. One of the reasons the installer doesn't offer it as an option is simple that no one has made the changes to the installer to make it do so. But, there are folks working on this. The ...


1

VMWare Workstation will certainly allow this. Add a hard drive and mark that you want to "Use a physical disk (for advanced users)". You can choose the entire disk or use individual partitions. Of course, nothing will allow both the host and the guest to mount the file system at the same time. That's a restriction of the file system, though, rather than a ...


1

For small-scale NAS amount of memory is more important than CPU power for ZFS, IMHO. Almost any x86 CPU on the market would do the job performance-wise, though some CPUs are better than others. You do want CPU that supports 64-bit mode. That rules out older Atoms. I would also recommend investing in a server board that supports ECC memory. ZFS's ability ...


1

For ZFS general recommendation is not to use hardware RAID controllers (or configure them in pass-through mode so that OS sees individual disks). ZFS takes care of redundancy and consistency checking. Alas ZFS is not that flexible. You may grow redundant zfs pool by replacing smaller drives in vdev with the larger ones one-by-one, but available space in ...


1

You assumption is correct. -r wouldn't have any effect in your case as there are no underlying file systems.


1

I ended up figuring out how to build a Gentoo install CD with ZFS support added: http://linux.arantius.com/building-a-gentoo-minimal-livecd-with-zfs-support This meant finding the Gentoo "releng" project, and updating it to add the ZFS tools and modules into the build: https://github.com/arantius/gentoo-releng/compare/master...zfs


1

3-drive RAIDZ will only give you 2x drive's peak throughput. So, ZFS array should be able to provide enough bandwidth to saturate GigE interface. The question is -- what serves that data over the network in your case? If it's samba, you do need to tweak it quite a bit in order to get decent throughput from it. In my case on FreeBSD, enabling aio and SMB2 ...


1

If you want to create a raid with zfs using different disk sizes you need to use "zpool create (name of your pool) raidz1 -f sdb sdc sdd" the -f arqument force zfs to use different sizes example 500gb 1tb 250gb hd


1

Don't have enough "reputation" to comment, so putting this in as a separate answer RE: Holger G suggestion while I agree that this is NOT safe... I found this thread looking for a way to go from RAIDZ1 to RAIDZ2 at a later date when extra drive can be acquired. So Holgers suggestion works great for me - create a RAIDZ2 with one dev as sparse file, offline ...


1

Also worth noting: on (at least) FreeBSD and Mac OS X, you can manipulate cpio files with tar. BSD tar uses libarchive under the hood, so it can handle cpio, pax, shar... This means that the usability issues of the cpio command doesn't have to stop you from interacting with cpio files.


1

The connection: I attempted to read the documentation on the Samba website, but was unable to determine whether Samba has error-correction. I had to assume the worst case - that Samba relies on the underlying network to be error free. If that underlying network is TCP/IP, it seems the only protection is a weak checksum. I wound up going with iSCSI because ...


1

Turns out the drive is failing. I did a scan using badblocks and it found tons of badblocks very quickly. Now I know what that looks like when the drive is part of a zpool.


1

If, as you said, you're not too worried about a potentially small failure rate then give ZFS with non-ECC a try. Set up the pool(s) with some form of redundancy (copies=2, raidz or raidz2) and check their integrity regularly with zpool scrub. Any bit errors will be picked up during the scrub and should be automatically repaired if the pool has some form of ...


1

Have a look to the zpool and zfs man pages. eg: zpool create mypool <disk or partition> zpool create mypool1 mirror <disk1> <disk2>


1

Unfortunately, ZFS-native encryption is closed-source. There are multiple options available: LUKS (other file-system-agnostic methods) File-based encryption on top of the FS, like ecryptfs. Setup is simple: Just do mount -t ecryptfs /tank/private /tank/private, it'll ask you a few questions and you're good to go. The Arch Linux Wiki contains a ...


1

Technically it doesn't really matter which filesystem you use. In practice there are some things to reconsider: You can only use ZFS on systems which have support for it. So you won't be able to mount your SD card on a Windows box. Linux boxes also take considerable effort. Embedded devices like camera's or media players don't support ZFS at all. Removing ...


1

In your setup it would not. What I would recommend is using ZFS built-in RAID-z2, which is then able to perform bit-rot recovery. Enabling lz4 compression will also provide better performance, as well as space savings.


1

I had the same Problem. Doing apt-get install ubuntu-zfs zfsutils and rebooting solved it for me. For some Reasons ubuntu did not update all zfs-packages. You can see if this is the case at the end of apt-get update output (packages held back)


1

As you said in the comments, you are using the Oracle ZFS Storage Appliance. This appliance is a demo of the (ex) Fishworks product and is meant to be used through the integrated CLI and web UI. You will not be able to use the standard ZFS commands (well, you could hack the appliance, but that's not what it was designed for). If you want to have a ...


1

After a looong debug session I found that the build and source links in /lib/modules/2.6.32-openvz-042stab090.5-amd64 to /usr/src and /usr/include were missing. Added them manually and it worked. Everyime happy that I can work in Linux where I can at least look into the source and debug...


1

Solaris 11 does not ask you if you want ZFS root because it has no other choice - only ZFS root is supported for Solaris 11 installs - UFS is relegated to supporting data disks only. See Oracle Solaris 11 File System Changes for more details.


1

You cannot make a snapshot of a zpool (zroot in your case). You can only take snapshots of filesystems on the zpool. What you can do is make snapshots of all relevant filesystems: for fs in $(zfs list | grep ^zroot | awk '{print $1;}' do echo "Making snapshot of ${i}" zfs snapshot ${fs}@fresh done


1

You are right your trouble is seek times. You should better use one of this two solutions : tar to create an archive of your dataset, I guess it will be faster or use snapshot directory function of ZFS



Only top voted, non community-wiki answers of a minimum length are eligible