block traversal size 819742315008 != alloc 1154012371456 (leaked 334270056448)

49 views
Skip to first unread message

Simon Casady

unread,
Nov 14, 2014, 3:28:21 PM11/14/14
to zfs-...@googlegroups.com
This is from zdb -b on a pool with one vdev of 2 2terabyte drives mirrored.

I can trash it and start over but is there away to "unleak" whats leaked ?
I will  recreate the pool anyway but would like to experiment with it for a while first.

A zfs recv failed with bus error 10 which apparently corrupted something.
zpool status shows no errors. However mount fails for some file systems.

BelecMartin

unread,
Nov 14, 2014, 4:06:14 PM11/14/14
to zfs-...@googlegroups.com
Do you have snapshots of what you consider 'leaked'?

You said a receive failed.... Have you checked your cables? Then shutdown and bring back online and get status.

I've seen lots of weird stuff and 90% can be resolved fairly quickly. You just won't be using every day ZPOOL commands but it is important to have a good understanding of your predicament and you really should be working on a bit for bit clone. 

I'll help in any way I can and you if you feel like poking the internals. 

Jason Belec
Sent from my "It's an iPod, a Phone, and an Internet Device..."
--

---
You received this message because you are subscribed to the Google Groups "zfs-macos" group.
To unsubscribe from this group and stop receiving emails from it, send an email to zfs-macos+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Simon Casady

unread,
Nov 14, 2014, 4:22:43 PM11/14/14
to zfs-...@googlegroups.com
These are internal drives and have never been flaky.  I use this pool as backup so nothing is lost I just have a need to understand  what may have happened before I recreate the pool and how, maybe, to recover.  The next time it may be more critical.
Looking at console there are crash dumps that show zfs crashing in avl_insert trying to update its mount table, not good.

I love poking internals. :)

BelecMartin

unread,
Nov 14, 2014, 4:31:50 PM11/14/14
to zfs-...@googlegroups.com
OK.

Of course 'never' is not a term associated with rotating rust. 

I have data that suggests 80% of failures are do to cables, they just take drives along for fun. Check the connectors, any split? Sometimes only noticeable when warm to hot. 

Have you check the power across various points on the supply, motherboard, drives?

If your really sure it cannot in any way be hardware, I'm at a loss. I have yet to see a single software related failure. 



Jason Belec
Sent from my "It's an iPod, a Phone, and an Internet Device..."

Simon Casady

unread,
Nov 15, 2014, 12:12:11 PM11/15/14
to zfs-...@googlegroups.com
Thanks for the comments Jason,  it apparently was a transient of some kind.  A scrub showed no problems but unmounts and mounts both failed and zdb -b showed a huge number of leaked space.

A reboot fixed everything.




BelecMartin

unread,
Nov 15, 2014, 12:24:30 PM11/15/14
to zfs-...@googlegroups.com
Awesome news!!! 


Jason Belec
Sent from my "It's an iPod, a Phone, and an Internet Device..."
Reply all
Reply to author
Forward
0 new messages