Memory Hogging in RHEL 4 Update 5

[raghuvendra]

Hi All,

I have a Linux based IVR running on a System that has 4GB of memory.
Max Simultaneous call capacity is 480. From TOP command i have noticied that
slowly total available is decreasing continuously. But if i look at the individual
usages (%MEM) process, its seems to be constant.

TOP OUTPUT:-

top - 13:54:17 up 12:36, 6 users, load average: 0.05, 0.10, 0.14
Tasks: 104 total, 1 running, 102 sleeping, 1 stopped, 0 zombie
Cpu(s): 2.5% us, 3.5% sy, 0.0% ni, 93.5% id, 0.5% wa, 0.0% hi, 0.0% si
Mem: 4153524k total, 3113276k used, 1040248k free, 180644k buffers
Swap: 7164904k total, 0k used, 7164904k free, 2704836k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
7607 root 16 0 241m 49m 6148 S 1 1.2 5:12.62 MGC-TCP
7659 root 15 0 181m 21m 4520 S 4 0.5 15:23.66 SdpMediadxp_New
5184 root 16 0 69932 16m 15m S 0 0.4 0:15.13 dlgcs7d
5585 root 18 0 16084 14m 14m S 0 0.4 0:00.04 gctload
5588 root 15 0 16064 14m 14m S 0 0.4 0:02.96 rsi
5683 root 15 0 16056 14m 14m S 0 0.4 0:09.65 rsi_lnk
5589 root 15 0 15912 14m 14m S 0 0.4 0:00.03 tim_lnx
5590 root 15 0 15912 14m 14m S 0 0.4 0:00.03 tick_lnx
7699 root 19 0 30948 6912 4976 S 0 0.2 0:32.95 OraApp
5088 root 16 0 120m 6736 1132 S 0 0.2 0:09.28 devmapserver
5747 root 16 0 9040 6388 1540 S 0 0.2 0:07.65 hald
5239 root 19 0 340m 5620 1528 S 0 0.1 0:00.58 dlgsysmonitorse
5307 root 17 0 108m 5296 1524 S 0 0.1 0:00.17 dlgsysclockserv
5562 root 16 0 308m 4884 1368 S 0 0.1 0:00.12 dlgclockdaemon
5155 root 18 0 108m 4560 1248 S 0 0.1 0:00.13 dlgtimeslotdole
5026 root 17 0 157m 4388 1068 S 0 0.1 0:00.12 OAMEventService
3138 root 16 0 9800 3472 1452 S 0 0.1 0:00.28 cupsd
5693 xfs 16 0 5972 3296 752 S 0 0.1 0:00.07 xfs
29897 root 16 0 10040 2940 1920 T 0 0.1 0:00.04 vim
3296 htt 17 0 8508 2936 2420 S 0 0.1 0:00.00 htt_server
18392 root 16 0 7668 2376 1856 S 0 0.1 0:05.79 sshd
29500 root 16 0 7896 2344 1856 S 0 0.1 0:15.29 sshd
28682 root 15 0 8256 2340 1860 S 1 0.1 0:27.37 sshd
30141 root 16 0 8544 2324 1856 S 0 0.1 0:07.01 sshd
23042 root 16 0 7028 2320 1860 S 0 0.1 0:00.03 sshd
1478 root 16 0 7980 2308 1856 S 0 0.1 0:00.08 sshd
3340 root 16 0 133m 2072 1200 S 0 0.0 0:02.84 RtfServer
29517 root 16 0 4592 1628 1200 S 0 0.0 0:00.22 bash
7711 root 16 0 57908 1500 1232 S 0 0.0 0:21.70 CurlAppMyStack
18406 root 15 0 5876 1492 1204 S 0 0.0 0:00.33 bash
1489 root 16 0 5372 1484 1200 S 0 0.0 0:00.09 bash
28726 root 15 0 4700 1484 1196 S 0 0.0 0:00.10 bash
30200 root 16 0 4644 1484 1200 S 0 0.0 0:00.07 bash
23164 root 15 0 5092 1440 1172 S 0 0.0 0:00.03 bash
7725 root 15 0 3600 1388 1012 S 0 0.0 1:47.97 SDP_52222
23202 root 16 0 4144 1148 912 S 0 0.0 0:00.00 curl
23205 root 16 0 4836 1148 912 S 0 0.0 0:00.00 curl


FREE OUTPUT:-

free -m
total used free shared buffers cached
Mem: 4056 3045 1010 0 176 2645

-/+ buffers/cache: 223 3833
Swap: 6996 0 6996

I have read that 223(Using Free) is the actual used and 2645 is cached that can taken be taken from linux in case if its requied. 223 is constant. But when i use the top command to see the total uses it is continously increasing (Currently, 3113276).

Please help. I am very worried.

Regards,
Raghuvendra Kumar

[coopstah13]

Linux memory management is different from windows pre-Vista. It will use all the available memory for caching to increase performance. As you noted, you are only actually using 223m. Top does not separate the cache memory from actual used memory, so that's why it seems different.