Scrubbing causes system lockup when running bitcoind

[lancethepants]

I'm running Debian Wheezy. I am using a 4 disk mirror configuration. I have one zfs file system that mounts to my home directory. I also have bitcoind running, which is stored in my home directory.
Whenever I run a scrub and bitcoind is running (daemonized), the whole system freezes up. The computer is neither accessible remotely, nor locally by ssh or login. I can still ping the box, but it will not serve anything. After a hard reboot, the scrub will successfully work when I kill bitcoind first. Tested this a couple times, and seems apparent to me that running bitcoind is causing the issue.

[AndCycle]

my configuration is 3 disk raidz,

I didn't get the lock up under Gentoo sys-fs/zfs-0.6.2-r2 with kernel 3.12.5,
but I do see bitcoind occupy 1st place in iotop,
so I move to current git version and that problem just gone,

I will suggest you tryout git head version, at least it works and haven't corrupt my entire system.

[behlendorf]

We try very hard to ensure master is always safe to run. That said, there is always the possibility the something will accidentally sneak in. But even in those case it's exceptionally unlikely to cause any persistent damage to the pool. Keeping your data safe is always job number one.

So I'd encouraging you also to try the latest source from Git. I think there's a pretty good chance it will resolve your issue.

[behlendorf]

This has been resolved in 0.6.3.