[OpenIndiana-discuss] Arc and L2arc hit rate

[jason matthews]

If you feel the need to micro manage the arc here are the parameters 
you'll like:
zfs_arc_max
zfs_arc_min

you need/should to specify values in 64 bits. ie.

* limit size of arc for hadoop systems
set zfs:zfs_arc_max = 0x900000000

or conversely
set zfs:zfs_arc_min = 0x800000000


understand that by nailing down the minimize size you are setting 
yourself up for memory starvation problems if you dont also tightly 
control memory usage. I dont use zfs_arc_min for this reason.

here is the oracle page on the tuneable, for that is worth.

http://docs.oracle.com/cd/E19253-01/817-0404/gjheb/index.html


j.

On 3/12/15 7:37 AM, Sebastian Gabler wrote:
> Hi,
>
> I have 12 GB RAM, 20 GB total pool size machine. L2ARC for one pool is 
> 107G SSD. I am wondering a bit about:
>
> - why not more of the RAM is used by ARC (I have run the ARC 
> optimisation script from the evil tuning guide, there machine is file 
> server exclusively, so no other apps steal RAM)
> - why the hit rate on the l2ARC is rather small (is it because the hit 
> rate on ARC is rather ok, which is what I think?)
> - would it help the quality of the service I am providing trying to 
> increase the prefetch hit rate, and how could I achieve that if.
>
> Sample from arcstat:
>
>     time hit% dh% ph% mh% l2hit% arcsz
> 14:20:52   92  99  27  99      1  4.5G 4.5G
> 14:21:52   92  99  28  99      1  4.5G 4.5G
> 14:22:52   92  99  31  99      1  4.7G 4.7G
> 14:23:52   93  99  30  99      1  4.5G 4.5G
> 14:24:52   92  99  27  99      1  4.5G 4.5G
> 14:25:52   92  99  30  99      1  4.5G 4.5G
> 14:26:53   92  99  27  97      1  4.5G 4.5G
> 14:27:53   93  99  30  99      1  4.5G 4.5G
> 14:28:53   92  99  28  99      1  4.5G 4.5G
> 14:29:53   93  99  29  99      1  4.5G 4.5G
> 14:30:53   92  99  28  99      1  4.5G 4.5G
> 14:31:53   92  99  30  99      1  4.5G 4.5G
> 14:32:53   93  98  30  99      2  4.5G 4.5G
> 14:33:53   92  99  28  99      1  4.5G 4.5G
> 14:34:53   92  99  27  99      1  4.5G 4.5G
> 14:35:53   92  99  32  99      1  4.6G 4.6G
> 14:36:53   89  98  32  99      0  4.5G 4.5G
> 14:37:53   92  99  30  99      1  4.5G 4.5G
> 14:38:53   92  99  28  99      1  4.5G 4.5G
> 14:39:53   93  99  28  98      2  4.5G 4.5G
> 14:40:53   93  99  29  99      1  4.5G 4.5G
>
> Thanks for any hints,
>
> Sebastian
>
> _______________________________________________
> openindiana-discuss mailing list
> openindiana-discuss at openindiana.org
> http://openindiana.org/mailman/listinfo/openindiana-discuss
>