[OpenIndiana-discuss] Crossbow performance & bit rot

[Jason Matthews]

So I have determined the trigger event for the "resets" in the response
times. It appears that there is some sort of memory leak in the kernel. When
memory utilization gets very high, I am not sure how high but it is around a
few hundred bytes on the freelists with zfs data near zero, the reset cycles
start on the vnics.

Page Summary                Pages                MB  %Tot
------------     ----------------  ----------------  ----
Kernel                    3806127             14867   91%
ZFS File Data                8870                34    0%
Anon                       277518              1084    7%
Exec and libs                5355                20    0%
Page cache                  13623                53    0%
Free (cachelist)            50984               199    1%
Free (freelist)             14275                55    0%

Total                     4176752             16315
Physical                  4176751             16315

The upside is, the memory leak eventually (which takes about two weeks to
get to this point after a reboot) makes the network response time stable
(but 10x higher than they should be) over extended periods of time. The
downside is, I have a memory leak that eventually has to be corrected with a
reboot. After the reboot, the reponse times destabilize requiring that the
VNICs be destroyed and re-created. It seems like the frying pan and fire...

I suspect excessive memory allocation and eventual starvation have a lot to
do with the weird symptoms. I wonder if the memory leak is an independent
process from the network issue or if they are directly related in some way. 


j.