Forum Discussion
Claud_13959
Nimbostratus
NimbostratusF5 memory usage
Just worried about high memory usage on my F5 8950, v11.1:
[root@f5:Active] config tmsh show sys memory | head
Sys::System Memory Information
------------------------------------------------------------------
Memory Used(bytes) Current Average Max(since 05/21/14 23:20:00)
------------------------------------------------------------------
Total Phys Memory 15.7G 15.7G 15.7G
OS Used Memory 15.4G 15.4G 15.4G
TMM Alloc Memory 8.2G 8.2G 8.2G
TMM Used Memory 954.3M 924.8M 954.8M
[root@f5:Active] config free -m
total used free shared buffers cached
Mem: 16078 15844 233 0 567 4195
-/+ buffers/cache: 11081 4997
Swap: 1023 7 1016
The TMM shows low usage but the Linux OS shows only 233MB free out of 15.7GB.
Here is a top:
top - 02:45:07 up 451 days, 4:34, 5 users, load average: 0.53, 0.34, 0.19
Tasks: 359 total, 4 running, 354 sleeping, 0 stopped, 1 zombie
Cpu(s): 7.3%us, 1.6%sy, 0.4%ni, 90.6%id, 0.2%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 16464436k total, 16233548k used, 230888k free, 581124k buffers
Swap: 1048504k total, 7448k used, 1041056k free, 4302064k cached
VIRT PID USER PR NI RES SHR S %CPU %MEM TIME+ COMMAND
1365m 15565 root RT 0 126m 125m S 12.9 0.8 893:22.50 tmm
1365m 15566 root RT 0 126m 125m S 10.6 0.8 717:47.65 tmm
1365m 15567 root RT 0 126m 125m S 8.9 0.8 810:56.81 tmm
1365m 15568 root RT 0 126m 125m R 7.3 0.8 955:00.22 tmm
1365m 15569 root RT 0 126m 125m S 6.3 0.8 723:16.05 tmm
1365m 15570 root RT 0 126m 125m S 6.6 0.8 768:51.00 tmm
1365m 15571 root RT 0 126m 125m S 6.0 0.8 606:04.78 tmm
1365m 15572 root RT 0 126m 125m S 7.6 0.8 621:15.98 tmm
445m 8188 mysql 20 0 66m 5468 S 0.0 0.4 125:29.68 mysqld
301m 8313 tomcat 20 0 251m 3016 S 0.0 1.6 302:25.53 java
181m 16014 root 20 0 30m 15m S 0.0 0.2 1:46.32 eam
171m 6643 root 20 0 39m 4036 S 0.0 0.2 80:17.65 java
138m 6450 root 20 0 97m 9.9m S 0.3 0.6 3022:29 mcpd
134m 15828 root 20 0 36m 18m S 0.0 0.2 4:04.94 apd
134m 5694 root 20 0 29m 11m S 0.7 0.2 4169:52 md
124m 16124 root 20 0 4676 3228 S 0.0 0.0 0:01.41 dpid
104m 6395 root 20 0 19m 9356 S 0.0 0.1 0:42.31 apmd
91992 6644 root 20 0 20m 7512 S 0.0 0.1 166:19.54 cbrd
87428 5750 root 20 0 9896 5788 S 0.0 0.1 28:12.64 chmand
80092 5646 root 20 0 15m 7476 S 0.0 0.1 5:09.39 monpd
75544 16263 root 20 0 23m 14m S 0.3 0.1 12:31.71 websso
75544 16292 root 20 0 23m 14m S 0.7 0.1 12:32.09 websso
75544 16349 root 20 0 23m 14m S 0.0 0.1 12:31.50 websso
74520 16169 root 20 0 23m 14m R 0.0 0.1 12:32.60 websso
74520 16276 root 20 0 23m 14m S 0.3 0.1 12:32.08 websso
74520 16305 root 20 0 23m 14m S 0.0 0.1 12:37.58 websso
74520 16319 root 20 0 23m 14m S 0.0 0.1 12:30.79 websso
74520 16333 root 20 0 23m 14m S 0.3 0.1 12:31.37 websso
Is this OK, or will the Linux OS run out of memory?
HHeredia_36237As long as i Know it's OK. Please note:
TMM Alloc Memory 8.2G 8.2G 8.2G TMM Used Memory 954.3M 924.8M 954.8M
TMM is using like 10-11% of it's allocated MEM.If this grows then you should be worried.
It's important to know the controller is a 'compound' of various subsystems and that tmm is the one you have to take care of.
However would be nice to wait for more comments on this because it's a very common issue among customers :)
hheredia
- What_Lies_Bene1
Cirrostratus
Amazing that this is still a topic of discussion after all these years.
First off, Claud, you may want to pull down my HMS book from here: https://www.wuala.com/sjiveson/Public/?key=BlKE7DHaCzCL to give you a better understanding of what the HMS is.
This article of mine explains how the HMS fits into the wider TMOS architecture: http://packetpushers.net/what-the-heck-is-f5-networks-tmos/.
Once you understand that, hopefully this will make sense;
- Any non-tmsh command you use in the HMS bash shell (like free) will generally only relate to the HMS.
- Top is one exception and will actually give you valid figures for CPU and memory (for the TMM processes), this is the only non-tmsh command you should rely upon.
- To avoid confusion it's best to simply rely on tmsh commands.
Claud_13959Thanks. This explains it.