Randy Bush wrote:
> where is exim_tidydb called? on freebsd 8-current, a number of systems
> running amd64 and i386, i have not been able to solve the following:
>
> Jun 6 00:20:06 work kernel: swap_pager_getswapspace(4): failed
> Jun 6 00:20:06 work kernel: swap_pager_getswapspace(16): failed
> Jun 6 00:20:06 work kernel: swap_pager_getswapspace(14): failed
> Jun 6 00:20:06 work kernel: swap_pager_getswapspace(16): failed
> Jun 6 00:20:06 work kernel: swap_pager_getswapspace(3): failed
> Jun 6 00:20:06 work kernel: swap_pager_getswapspace(16): failed
> Jun 6 00:20:06 work kernel: swap_pager_getswapspace(16): failed
> Jun 6 00:20:06 work kernel: pid 12159 (exim_tidydb), uid 0, was killed: out of swap space
>
> and i was looking for where exim_tidydb was run to see if commenting it
> out would move the problem. i assumed exicyclog, but if it is there it
> is too subtle for first cuppa.
>
> randy
>
/usr/local/etc/periodic/daily/150.exim-tidydb
But IMNSHO, running out of swap space calls for providing more swap space
(and/or RAM).
Two or more small swaps per physical device 'spindle' helps reduce head movement
as *BSD will use the 'closest' one(s) automagically.
Killing off tidydb could be expected to make the problem worse over time, not
better, as an ever-growing DB size has to be waded through.
Bill Hacker
