It has been my experience, that in a system with a lot of cache (meaning the VM guest has too much memory allocated to it for the workload being run, in most cases) that when you need memory, you will get it from cache before you swap to disk. I tested this several times.
I also found a command that allows you to flush the buffers to disk, in
SLES8 called flushb.
I created a script to flush a small 2 disk system
#!/bin/sh
# Flush the buffers for /dev/dasdb1 and /dev/dasdc1
echo "Buffers before flush"
cat /proc/meminfo | grep Buffers
/usr/sbin/flushb /dev/dasdb1
/usr/sbin/flushb /dev/dasdc1
echo "Buffers after flush"
cat /proc/meminfo | grep Buffers
"Levy, Alan"
<[EMAIL PROTECTED]
gov> To
Sent by: Linux on [EMAIL PROTECTED]
390 Port cc
<[EMAIL PROTECTED]
IST.EDU> Subject
Recycling servers ?
09/03/2004 12:54
PM
Please respond to
Linux on 390 Port
<[EMAIL PROTECTED]
IST.EDU>
I currently have a few linux servers running in production (ZVM 4.4)
with heavy usage. Our IFL is only IPL'd once every few months. Would it
be advisable to recycle the whole machine (and all linux servers running
under VM) more than that ? I was told that linux servers have to be
rebooted periodically to clear out buffers, cache, storage, etc.
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or
visit
http://www.marist.edu/htbin/wlvindex?LINUX-390
(Embedded image moved to file: pic32029.jpg)
----------------------------------------------------------------------
For LINUX-390 subscribe / signoff / archive access instructions,
send email to [EMAIL PROTECTED] with the message: INFO LINUX-390 or visit
http://www.marist.edu/htbin/wlvindex?LINUX-390<<attachment: pic32029.jpg>>
