Size of /var/log/messages vs. Size of cached - Embedded
This is a discussion on Size of /var/log/messages vs. Size of cached - Embedded ; Is there a correlation between the size of the /var/log/messages file
and the amount reported as "cached" when the top utility is run? I
see that if I put a lot of log messages in the /var/log/messages file
that the ...
-
Size of /var/log/messages vs. Size of cached
Is there a correlation between the size of the /var/log/messages file
and the amount reported as "cached" when the top utility is run? I
see that if I put a lot of log messages in the /var/log/messages file
that the amount cached directly increases. If I then remove the /var/
log/messages file, the amount cached does not decrease.
Is this the expected behavior? In general, is there a way to keep the
amount cached down? I notice my system crashing when the amount
cached reaches a certain amount.
I am running Linux 2.6.10 on PowerPC.
-
Re: Size of /var/log/messages vs. Size of cached
Bill writes:
>Is there a correlation between the size of the /var/log/messages file
>and the amount reported as "cached" when the top utility is run? I
>see that if I put a lot of log messages in the /var/log/messages file
>that the amount cached directly increases.
"Cached" refers to file data that's cached in RAM, so if you write a
lot to a file, the cached file data will increase (unless there is
something else competing for RAM). So yes, writing to
/var/log/messages will tend to increase the amount cached; but
normally the amount written to /var/log/messages is small compared to
the cache size.
>If I then remove the /var/
>log/messages file, the amount cached does not decrease.
Probably the syslogd still has the file open, so it's not really
deleted, just unlinked.
>Is this the expected behavior? In general, is there a way to keep the
>amount cached down?
Start a memory-intensive application that does not access files.
> I notice my system crashing when the amount
>cached reaches a certain amount.
That's a bug in your kernel. As others have mentioned, try something
more recent if possible.
- anton
--
M. Anton Ertl Some things have to be seen to be believed
anton@mips.complang.tuwien.ac.at Most things have to be believed to be seen
http://www.complang.tuwien.ac.at/anton/home.html