Running out of swap space????

[Dan Nelson]

In the last episode (Jun 05), pauls at utdallas.edu said:
> I've got a server that is running out of swap space:
> 
> +pid 37308 (mysqld), uid 88, was killed: out of swap space
> +swap_pager: out of swap space
> +swap_pager_getswapspace(1): failed
> 
> The strange this is, this server has a 6GB swap partition!
> 
> swapinfo -h
> Device          1K-blocks     Used    Avail Capacity
> /dev/da0s1b       6291456     2.6G     6.0G    43%
> 
> This isn't exactly a resource-starved machine either:
> 
> real memory  = 2146893824 (2047 MB)
> avail memory = 2065797120 (1970 MB)
> 
> last pid: 52327;  load averages:  0.45,  0.46,  0.45 
> up 11+03:42:04  03:32:15
> 63 processes:  1 running, 62 sleeping
> CPU states:  5.3% user,  0.0% nice,  0.8% system,  3.9% interrupt, 90.1% idle
> Mem: 1410M Active, 126M Inact, 190M Wired, 82M Cache, 214M Buf, 78M Free
> Swap: 6144M Total, 2687M Used, 3457M Free, 43% Inuse

Seems sort of starved to me; 2GB of RAM yet you have enough processes
active to have allocated all of that plus 2.6GB of swap!  I wouldn't be
surprised if occasionally you allocated another 3GB, which would max
out your swap space.  With only 120 processes total, you can probably
just run top sorted by size (top -o size, or enter "osize" when it's
running) and look for large ones.

-- 
	Dan Nelson
	dnelson at allantgroup.com