Slow shutdown with big file in /dev/shm

[Bojan Smojver]

Adam Jackson <ajax at redhat.com> wrote:

>Presumably because that file got paged out.
>
>Think about it.  Disabling swap doesn't know that the file exists only 
>on a RAM-based file system, and even if it did, doesn't know that we're
>
>about to shut down.  So swapoff has to assume that any pages currently 
>in the swap area being disabled might be valuable, and has to read them
>
>back into memory.
>
>Now as to why we disable swap on shutdown, I'm not really sure.  It 
>certainly seems like useless work to me.

Would it help if all memory based file systems got unmounted before swap gets disabled? That would kill these files, right?

-- 
Bojan