fsck with ZFS ?

[mikedeissler]

I've been trying out the nightly builds of version 8. I'm new to ZFS. I notice there doesn't seem to be a way to do a disk check in the GUI, unless I'm stumbling by it.

I'm log in via SSH (as a wheel member) and tried to umount the drive (even tried -f) but the "operation is not permitted."

1st question is, does ZFS need periodic fscks or does the system take care of it on it's own somehow? And, if it does not and a periodic fsck is good practice, how do I do it ?

The new GUI rocks.

Thx, mik

 

[iJimmyC]

Hey pal, I'm new here (joined today..) so I'm not entirely sure what you're doing...

However, I believe that there is no fsck in ZFS.

http://blogs.oracle.com/video/entry/becoming_a_zfs_ninja

This is a link to a couple of vids from a guy who a) knows his S**T and b) makes a seemingly boring subject funny..

If you look about 6 minutes in you see that the structure of ZFS means that there is no fsck... - even if this isn't useful in this instance I think the vids help to further understand the structure etc...

I'm also including this link that I've come across on the freenas doc that may help if you're after checking disk integrity...

http://doc.freenas.org/index.php/S.M.A.R.T._Tests

Hope this helps and is what you were looking for...

 

[ProtoSD]

Take a look at 'zpool scrub'...

 

[mikedeissler]

Hey pal, I'm new here (joined today..) so I'm not entirely sure what you're doing...

However, I believe that there is no fsck in ZFS.

http://blogs.oracle.com/video/entry/becoming_a_zfs_ninja

This is a link to a couple of vids from a guy who a) knows his S**T and b) makes a seemingly boring subject funny..

If you look about 6 minutes in you see that the structure of ZFS means that there is no fsck... - even if this isn't useful in this instance I think the vids help to further understand the structure etc...

I'm also including this link that I've come across on the freenas doc that may help if you're after checking disk integrity...

http://doc.freenas.org/index.php/S.M.A.R.T._Tests

Hope this helps and is what you were looking for...

Thanks for the amazing link - had no idea ZFS was such a departure from the ways of the Ancients - mik

 

[mikedeissler]

Take a look at 'zpool scrub'...

Thanks Protosd - Exactly what I needed. Mik

 

[jgreco]

Filesystem technologies seem to stagnate for years at a time in many operating systems. This is both good and bad, good because we want to be able to do upgrades, maintain existing files, and of course the filesystems are one of the most important bits of technology, so you want them reliable (see fun with Linux Ext2, etc., historically). On the other hand, the vast amounts of data we can now store changes things substantially. It used to be a thousands-of-dollars affair to add a hard drive back in the early FFS/UFS days, and storage was measured in megabytes. Now people want to be able to add a hundred dollar 3TB SATA drive to their system (and yes NewEgg was recently selling them for $109 with coupon code), CPU's and memory are super fast and super cheap, and we should be able to abstract things so much further than we were doing 20-30 years ago.

ZFS represents a vision for a more advanced method to store files, and while it misses the target in some ways, it's a competent system that brings us closer to something for this century. Theoretically one shouldn't need a tool to check the consistency of your filesystem, if you always leave it in a usable state, though one could argue that a tool that was a fs_CK_er and not a filesystem FIXER might be nice for us paranoids, because it isn't entirely clear that 'scrub' is sufficient.

 

[mikedeissler]

To be sure...

Filesystem technologies seem to stagnate for years at a time in many operating systems. This is both good and bad, good because we want to be able to do upgrades, maintain existing files, and of course the filesystems are one of the most important bits of technology, so you want them reliable (see fun with Linux Ext2, etc., historically). On the other hand, the vast amounts of data we can now store changes things substantially. It used to be a thousands-of-dollars affair to add a hard drive back in the early FFS/UFS days, and storage was measured in megabytes. Now people want to be able to add a hundred dollar 3TB SATA drive to their system (and yes NewEgg was recently selling them for $109 with coupon code), CPU's and memory are super fast and super cheap, and we should be able to abstract things so much further than we were doing 20-30 years ago.

ZFS represents a vision for a more advanced method to store files, and while it misses the target in some ways, it's a competent system that brings us closer to something for this century. Theoretically one shouldn't need a tool to check the consistency of your filesystem, if you always leave it in a usable state, though one could argue that a tool that was a fs_CK_er and not a filesystem FIXER might be nice for us paranoids, because it isn't entirely clear that 'scrub' is sufficient.

You're so right jgreco - anything that moves can break, electrons included....