SUMMARY: strange vmstat stats

From: Caparrosso, Nelson T. (Nelson.T.Caparrosso@Mail.AAS.ameritech.com)
Date: Thu Aug 10 2000 - 13:37:36 CDT


Thanks to everyone that made quick responses (too many to mention). Original
post below.

I know that vmstat output are just statistics and the very first line would
tend to be useless as it is supposed to be the average since uptime and
somehow since the server's been up for quite some time, the stats must have
flipped. I just could not get an explanation of the negative numbers. The
continued high values in 'proc w' will not be all that strange as it shows
that there were episodes that the server's had memory deficiencies.

A very good explanation from Bill Hathaway [wdh@perfectorder.com] about this
scenario follows:
--------------------------------
Nelson,
  the reason you are getting some negative values for the first line is
that the first line is a summary of average since boot. If your machine
has been up for a while, the counters may have flipped. Always throw out
the first line when looking at vmstat output.
 The w column means "runnable but swapped" which probably occured at some
point when the system was low on memory, some of the processes that had
not ran in a while were paged out, and they have never had a need to run
again, so they stayed swapped out. A good example of this would be
lpsched on a system that never did any printing. There is not necessarily
a bad thing, and paging the processes back in before they try to execute
again would be a waste of system resources. As far as I know, there is
not a way to reset the vmstat counters ( which are actually just kernel
statistics), if you don't like the vmstat output, you can always roll your
own using Adrian Cockroft's SE toolkit which has a C like language for
getting at misc. kernel statistics.
/wdh
--------------------------------

Original Post:
---------------------------
Esteemmed Gurus:

I'd appreciate very much if anyone could help me figure out the following
vmstat stats:

 r b w swap free si so pi po fr de sr s5 s6 s7 s1 in sy cs us sy
id
 0 0 13 5672 2288 0 0 0 0 0 0 0 0 6 1 0 4294967196 0 0 -15
-12 -139
 0 0 36 2781720 127568 0 0 0 0 22 0 20 0 9 0 0 315 9596 903 15 7
78
 0 0 36 2779544 125912 0 0 0 0 198 0 205 0 3 1 0 256 8216 995 19 6
75
 0 0 36 2776632 124448 0 0 0 0 372 0 344 0 1 0 0 250 8493 1021 19 6
75
 0 0 36 2776960 126088 0 0 0 0 248 0 203 0 3 1 0 357 11142 1303 21 7
72

Server is an UE 6000 (10x400Mhz/8MbEC), 4 GB memory, 4 GB swap, Oracle
database, Solaris 2.6 with the freshest patches possible. Server's been up
for over a month and have had several episodes of very heavy processing and
I/O. I am just baffled by the first line which shows negative stats and the
succeeding 'procs w' stats that show a high number of process waits. My ps
show not much processes that are waiting and my VM is okay.... I there a way
to reset VMSTAT so it shows fresh stats?

TIA and I will summarize.

S
U BEFORE POSTING please READ the FAQ located at
N ftp://ftp.cs.toronto.edu/pub/jdd/sun-managers/faq
. and the list POLICY statement located at
M ftp://ftp.cs.toronto.edu/pub/jdd/sun-managers/policy
A To submit questions/summaries to this list send your email message to:
N sun-managers@ececs.uc.edu
A To unsubscribe from this list please send an email message to:
G majordomo@sunmanagers.ececs.uc.edu
E and in the BODY type:
R unsubscribe sun-managers
S Or
. unsubscribe sun-managers original@subscription.address
L To view an archive of this list please visit:
I http://www.latech.edu/sunman.html
S
T



This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:14:14 CDT