top和vmstat、prtstat如何准确定位内存使用情况
在我的M5000上，用top和vmstat看到空闲内存一般是在2.5G，偶尔空闲内存是20多G,我现在很想找到谁占用那么大的内存；更重要的是看不出来top怎么样把内存使用情况都输出，我视图用top -n 2000,汇总size值
也远低于使用的内存；我用prstat看很多进程使用近10G的RSS（常驻内存）；还有prstat后面好像统计出每个用户使用内存情况和实际好像差的也很大。
1、我用top看到：
Memory: 64.0G real, 2.5G free, 10.3G swap in use, 78.9G swap free
PID USERNAME THR PR NCESIZE RES STATE TIME FLTS CPU COMMAND
6765 mqadmin 52 32 0539M517M cpu31 231.6H 0 10.96% dispatcher
9333 mqadmin 49 55 0411M390M sleep 452.3H 09.23% dispatcher
12361 mqadmin 50 31 0225M202M cpu0642.3H 09.20% dispatcher
13487 mqadmin 50 52 0369M347M sleep32.5H 08.52% dispatcher
28215 mqadmin 48 49 0307M285M sleep97.5H 05.18% dispatcher
5656 mqadmin 20 34 0 89.9M 66.1M sleep 196.2H 00.66% dispatcher
11521 mqadmin 3 59 0 43.7M 29.2M sleep87.7H 00.59% dbreoperate
23660 mqm 66 59 0 0K 0K sleep 272.5H 00.39% amqzlaa0_nd
5563 mqadmin 11 49 0 14.2M 8080K sleep82.2H 00.35% transponder
29652 mqm 33 59 0 0K 0K sleep 301.8H 00.27% amqrmppa
15701 mqm 33 59 0 0K 0K sleep 233.4H 00.27% amqrmppa
3180 mqm 36 59 0 0K 0K sleep 277.5H 00.22% amqrmppa
4788 mqm 64 59 0 0K 0K sleep 218.0H 00.20% amqzlaa0_nd
2、vmstat 看到空闲的内存一般情况也在2左右：
kthr memory page disk faults cpu
r b w swapfreeremf pi po fr de sr m0 m1 m3 m4 in sy cs us sy id
3 0 0 82767088 24325064（23G） 38 246 0 0 0 000 4800 189372 449823 290124 16 16 68
13 0 0 61068120 2663544 （2.5G）3 41 0 00000000 238216 540095 363472 20 20 60
1 0 0 61067696 2663120（2.5G） 2 17 000000000 240154 516411 361832 19 20 61
3 0 0 61067696 2663120（2.5G） 40 356 0 0 0000000 331132 736145 516640 27 27 46
3、下面是prstat的输出：
Total: 137 processes, 2829 lwps, load averages: 37.48, 36.88, 28.84
PID USERNAMESIZE RSS STATEPRI NICE TIMECPU PROCESS/NLWP
6765 mqadmin 539M517M cpu17 51 0 231:36:4610% dispatcher/52
9333 mqadmin 411M390M cpu23 50 0 452:17:51 8.1% dispatcher/49
12361 mqadmin 225M202M cpu17 41 042:15:13 7.9% dispatcher/50
13487 mqadmin 369M347M cpu1 32 032:30:14 7.3% dispatcher/50
28215 mqadmin 307M285M sleep 41 097:29:40 4.4% dispatcher/48
5656 mqadmin 90M 66M cpu49 39 0 196:09:31 0.6% dispatcher/20
11521 mqadmin 44M 29M sleep 59 087:44:30 0.6% dbreoperate/3
5563 mqadmin 14M 8080K sleep 54 082:12:05 0.4% transponder/11
23660 mqm 9387M 8476M sleep 59 0 272:32:05 0.4% amqzlaa0_nd/66
29652 mqm 77M 57M sleep 59 0 301:45:06 0.3% amqrmppa/33
15701 mqm 77M 58M sleep 59 0 233:24:57 0.3% amqrmppa/33
7411 mqadmin3968K 3416K cpu21 59 0 0:00:05 0.2% prstat/1
3180 mqm 77M 57M sleep 59 0 277:29:09 0.2% amqrmppa/36
5645 mqadmin 40M 21M cpu4 59 052:44:37 0.2% dispatcher/4
4788 mqm 9388M 8833M sleep 59 0 217:58:26 0.2% amqzlaa0_nd/65
6075 mqm 77M 58M sleep 59 0 258:42:54 0.2% amqrmppa/42
18359 mqm 77M 58M sleep 59 0 162:38:47 0.2% amqrmppa/43
15704 mqm 77M 57M sleep 59 0 169:40:46 0.2% amqrmppa/37
5573 mqadmin 39M 24M sleep 52 044:55:08 0.2% dispatcher/4
5360 mqm 9390M 8801M sleep 59 0 275:20:50 0.2% amqzlaa0_nd/56
19697 mqm 77M 58M sleep 59 0 190:35:29 0.2% amqrmppa/39
22631 mqm 77M 58M sleep 59 0 304:30:17 0.2% amqrmppa/34
29908 mqm 9389M 8390M sleep 59 088:25:28 0.2% amqzlaa0_nd/66
6106 mqm 9391M 8762M cpu56 59 0 244:19:15 0.1% amqzlaa0_nd/66
1596 mqm 77M 57M sleep 59 0 243:50:10 0.1% amqrmppa/38
18822 mqm 77M 57M sleep 59 0 179:14:13 0.1% amqrmppa/36
16645 mqm 9395M 8818M sleep 59 0 125:45:08 0.1% amqzlaa0_nd/66
15691 mqm 77M 57M sleep 59 0 219:45:08 0.1% amqrmppa/41
23670 mqm 9388M 8840M sleep 59 0 269:41:48 0.1% amqzlaa0_nd/61
15768 mqm 77M 58M sleep 59 0 272:42:04 0.1% amqrmppa/39
5190 mqm 77M 58M sleep 59 0 224:00:11 0.1% amqrmppa/33
19416 mqm 9389M 8450M sleep 59 0 185:34:14 0.1% amqzlaa0_nd/58
22457 mqm 9391M 8874M sleep 59 0 201:01:16 0.1% amqzlaa0_nd/64
NPROC USERNAMESWAP RSS MEMORY TIMECPU
31 mqadmin1950M 1932M 2.9% 1400:15:340%
67 mqm 9924M 9799M 15% 9513:40:0 5.0%
5 patrol 22M 32M 0.0%79:17:07 0.0%
2 nobody 49M 62M 0.1% 3:33:44 0.0%
1 smmsp 2464K 10M 0.0% 0:01:11 0.0%
1 daemon 2856K 7312K 0.0% 0:04:48 0.0%
2 dwboco 552K 3696K 0.0% 0:00:00 0.0%
1 noaccess176M153M 0.2% 1:47:10 0.0%
27 root 76M 88M 0.1% 2:02:16 0.0%
页:
[1]