I have a Alpha 2100 4/275 running dUnix 3.2c. The machine has 500M
physical memory, 400M swap, swap deferred mode. A user is trying to access
all the memory possible for a large model, and is unable to account for
memory usage on the machine.
o total active 64.4M from 'ps vxa' (RSZ below)
o vmstat shows active of 24 8-K pages (192M)
. vmstat also shows 33 8-K pages free memory plus 4.8 8-K wired (302.4M)
Note that vmstat reports a much higher active memory total than the process
listing (ps vxa) shows (~130M discrepancy). The user WANTS those 130M of
memory and I can't tell him where they are. Can anyone help? I need a lucid
description of where that memory is. It is a heavily used machine and I am
unlikely to get time to shut it down and watch available memory with only
the single user. A larger swap partition would also be fairly difficult
right now.
--Grant
Notes:
vmstat DOES show that wired+free+active ~= 500M, the correct total.
Swap is less than physical memory. 400M of swap space in deferred mode.
Total VSZ 831.8M from 'ps vxa' (below).
--->ps vxa ----------------------
PID TTY S TIME SL PAGEIN VSZ RSS %CPU %MEM COMMAND
* 0 ?? R < 01:42:36 5129891 77 522M 202 0.0 4.0 [kern..
+13414 ttyp7 R N 1-03:20:59 0 0 18.3M 143 99.0 2.7 bcl
8539 ttyp4 R N 04:13:27 0 0 9.32M 5.9M 99.0 1.2 micom
11787 ttyp2 R N 02:02:15 0 0 8.24M 3.3M 99.0 0.6 sillco
8409 ?? I 0:02.93 37 61 8.64M 2.0M 0.0 0.4 /usr/b
....
* I assumed RSS was 20.2M for this process. I included the
RSS and VSZ for [kern idle] in the totals above.
+ I assumed RSS 14.3M for totals.
--
Grant Basham (305)361-4026 University of Miami
grant_at_oj.rsmas.miami.edu RSMAS Computer Facility/Systems
Received on Tue Mar 12 1996 - 23:06:13 NZDT