=?unknown-8bit?B?WmFuZfjhZOxu4SBwYW3s?= =?unknown-8bit?Q?=BB?=

[Jozef Babjak]

> zacatek staci TOP a zjistit, kdo vlastne ma tu pamet na svedomi.

  ^-- Stale ale narazam na problem, ako udaje zobrazovane v top vlastne
interpretovat. Napr. u mna teraz zobrazuje toto:

last pid: 88929;  load averages:  2.51,  2.56,  2.57    up 0+23:00:16  13:57:51
99 processes:  3 running, 95 sleeping, 1 zombie
CPU states: 85.2% user,  0.0% nice, 14.5% system,  0.4% interrupt,  0.0% idle
Mem: 338M Active, 749M Inact, 188M Wired, 53M Cache, 112M Buf, 175M Free
Swap: 1024M Total, 1024M Free

  PID USERNAME    THR PRI NICE   SIZE    RES STATE    TIME   WCPU COMMAND
  663 root          1  76    0   174M   136M select  19:02  0.00% Xorg
 6894 babjak        4  20    0   108M 84732K kserel  14:46  0.00% firefox-bin
19705 root          1   8    0 41240K 40204K wait     0:31  0.00% ruby18
 1239 babjak        1  76    0 31560K 18180K select   0:01  0.00% kdeinit


Este zlozitejsie to vidim napr. u Solarisu, kde neviem ani poriadne 
dat dohromady udaje z top  a  prstat. Z toho isteho servera v tej istej
chvili zobrany top:

last pid:  3786;  load averages:  0.13,  0.12,  0.11                   14:01:50
147 processes: 145 sleeping, 1 zombie, 1 on cpu
CPU states: 92.7% idle,  0.8% user,  6.5% kernel,  0.0% iowait,  0.0% swap
Memory: 2048M real, 373M free, 2218M swap in use, 3189M swap free

   PID USERNAME LWP PRI NICE  SIZE   RES STATE    TIME    CPU COMMAND
  3786 root       1  59    0 2232K 1792K cpu/1    0:00  0.83% top
 14972 root      47  59    0  241M   68M sleep  211:13  0.23% java
  3744 root       1  59    0 4912K 4632K sleep    0:00  0.19% prstat
 29352 tomcat    48  59    0  349M  303M sleep   21:22  0.17% java
  2619 mailsrv    3  59    0  181M   20M sleep  130:20  0.09% mshttpd
   556 ldap      46  59    0  157M   15M sleep  643:27  0.07% ns-slapd
  1279 icsuser    9  59   -2   55M   17M sleep  368:56  0.06% csadmind
   654 root      70  59    0  100M 9608K sleep  308:45  0.05% ns-httpd
   963 webservd 146  59    0  349M   92M sleep  299:18  0.05% webservd
 13563 oracle     1  59    0  378M  353M sleep   48:59  0.05% oracle
   932 root      33  59    0  151M   29M sleep  257:17  0.04% webservd
 15007 root      39  59    0  163M   46M sleep   34:10  0.03% java
  1315 icsuser    1  59   -2 4704K 1584K sleep   44:35  0.03% csstored.pl
 13545 oracle     1  59    0  379M  353M sleep   32:05  0.03% oracle
 13555 oracle    11  59    0  380M  353M sleep   22:55  0.02% oracle
  6873 svn        1  59    0   23M 3472K sleep    9:18  0.02% httpd
  2617 mailsrv    3  59    0  155M   13M sleep   45:13  0.02% imapd
 25621 root       1  59    0   15M 2432K sleep    0:55  0.02% smbd

a prstt -a:

   PID USERNAME  SIZE   RSS STATE  PRI NICE      TIME  CPU PROCESS/NLWP
  3786 root     2232K 1792K sleep   59    0   0:00:00 0.8% top/1
  3744 root     4912K 4632K cpu0    59    0   0:00:00 0.2% prstat/1
 14972 root      241M   68M sleep   59    0   3:31:13 0.2% java/47
 29352 tomcat    349M  303M sleep   59    0   0:21:22 0.2% java/48
  2619 mailsrv   181M   20M sleep   59    0   2:10:20 0.1% mshttpd/3
   556 ldap      157M   15M sleep   59    0  10:43:27 0.1% ns-slapd/46
  1279 icsuser    55M   17M sleep   59   -2   6:08:56 0.1% csadmind/9
   963 webservd  349M   92M sleep   59    0   4:59:18 0.0% webservd/146
   654 root      100M 9608K sleep   59    0   5:08:45 0.0% ns-httpd/70
 13563 oracle    378M  353M sleep   59    0   0:48:59 0.0% oracle/1
   932 root      151M   29M sleep   59    0   4:17:17 0.0% webservd/33
 15007 root      163M   46M sleep   59    0   0:34:10 0.0% java/39
  1315 icsuser  4704K 1584K sleep   59   -2   0:44:35 0.0% csstored.pl/1
 13545 oracle    379M  353M sleep   59    0   0:32:05 0.0% oracle/1
  6873 svn        23M 3472K sleep   59    0   0:09:18 0.0% httpd/1
 NPROC USERNAME  SIZE   RSS MEMORY      TIME  CPU
    77 root     1220M  234M   2.9%  21:36:09 1.6%
     1 tomcat    349M  303M   3.8%   0:21:22 0.2%
    11 mailsrv   716M  122M   1.5%   3:37:16 0.1%
    21 oracle   7612M 7065M    88%   3:13:29 0.1%
     6 icsuser   264M   47M   0.6%   9:00:40 0.1%
     1 ldap      157M   15M   0.2%  10:43:27 0.1%
Total: 148 processes, 897 lwps, load averages: 0.13, 0.12, 0.11

Co mi pride najdivnejsie, ak spocitam polozky "rss" alebo polozky "size", 
dojdem k suctu daleko vyssiemu, ako ram+swap: ten server ma 2GB ram a 
-- podla vsetkeho -- 3.1GB swap, aj ked netusim, ako admin prave 
k takejto hodnote prisiel. Pricinou moze byt shared memory, ktora 
sa zapocitava kazdemu procesu, ktory ju vyuziva; preto tolko pamate
pre oracle. 


Ako teda interpretovat udaje zobrazene programom top(1) vo FreeBSD?
Ako ich interpretovat vo vseobecnosti na inych unixovskych systemoch?

J.