[zfs-discuss] Crash bug related to cache

Matthew Robbetts wingfeathera at gmail.com
Thu Feb 16 11:47:14 EST 2012


On 16 Feb 2012, at 15:11, Manuel Amador wrote:

> Please don't thread-ride.  As for the bug, you need to limit your zfs_arc_max 
> to less than half the RAM on your system to solve this problem, or update to 
> the latest master HEAD and use that.
> 

Thread-ride? I'm sorry, I don't know what that is. As I said, I can't find an issue on the bug tracker, so I thought it would be reasonable to ask on the list. Where else would I ask? And if netiquette is an issue, I hear top-posting annoys some people :p

Anyway, I've done both of what you suggest. 2GB RAM with a 512MB ARC limit:

cat `find /sys -name zfs_arc_max`
536870912

And I update to latest code quite regularly. No luck.

As soon as the normal linux page cache grows to fill up the remainder of the RAM, and I then start anything involving ZFS (scrub, rsync, etc), the system reliably locks up within a few seconds.


> 
> On Thursday, February 16, 2012 14:42:20 Matthew Robbetts wrote:
>> Hi all,
>> 
>> I'm just wondering if there's any news with regard to the crash bug that is
>> allayed by looping 'echo 3 > /proc/sys/vm/drop_caches'. I can't seem to
>> find an entry on the bugtracker, but I remember others besides me have
>> reported seeing it before.
>> 
>> Does anyone know why it happens? This makes ZoL unusable for me.
>> 
>> Thanks,
>> Matt



More information about the zfs-discuss mailing list