Performance system

[Hunan_1]

Hi all,

May someone tell me if the memory assigned to my HP servers is OK and how's the performance.

These are the messages displayed when I run the script which code source is in the file attached (I downloaded this from ITRC).

Furthermore I would really please if you could explain me what the fields shown mean.

1.-
Memory Stat total used avail %used
physical 1024.0 647.8 376.2 63%
active virtual 334.5 218.6 115.9 65%
active real 356.4 240.1 116.3 67%
memory swap 0.0 0.0 0.0 ?%
device swap 1024.0 488.8 535.2 48%
2.-
Memory Stat total used avail %used
physical 512.0 493.7 18.3 96%
active virtual 261.4 74.4 187.0 28%
active real 293.4 65.9 227.5 22%
memory swap 358.1 88.0 270.1 25%
device swap 1024.0 513.4 510.6 50%
3.-
Memory Stat total used avail %used
physical 1536.0 572.0 964.0 37%
active virtual 119.2 20.0 99.3 17%
active real 158.3 11.7 146.6 7%
memory swap 1133.8 169.6 964.2 15%
device swap 1024.0 223.4 800.6 22%
4.-
Memory Stat total used avail %used
physical 2048.0 1868.9 179.1 91%
active virtual 311.2 60.0 251.2 19%
active real 139.4 20.7 118.8 15%
memory swap 1579.2 1163.4 415.8 74%
device swap 1024.0 1024.0 0.0 100%

Thanks

[Steve Steel · Accepted solution]

Hi

memory wise the performance is ok

Memory Stat total used avail %used
physical 2048.0 1868.9 179.1 91%
active virtual 311.2 60.0 251.2 19%
active real 139.4 20.7 118.8 15%
memory swap 1579.2 1163.4 415.8 74%
device swap 1024.0 1024.0 0.0 100%

This is a programmed version of swapinfo -at

Physical memory = 2gb 91% used . Fine
Active virtual = 19% used fine
Active real = 15% fine
memory swap
Based on dbc_max_pct and dbc_min_pct in the kernel which allocates dynamic buffer caching in place of using the device swap.This is where most of your memory has gone sice 1gb or 50 is reserved or used for swapping.dbc_max_pct = 50% by default.

You are not using it all so no problem.

Device swap is unused so in case of memory shortages put dbc_max_pct on 30% and it will device swap.

See man swapinfo and www.docs.hp.com for more info

Steve Steel

[Ravi_8]

Hi

details that u have furnished shows that your system is healthy.

[Stefan Farrelly]

The first 3 of your servers look fine. The 4th one doesnt.

All show physical memory < 100% (its fine to run at 90 odd %, just no 100)
Memory swap is on which is a safety net.
Device swap is well < 100% on the first 3, which is good, but the 4th server shows device swap at 100% which means at some point you have run out of physical memory (even though physical memory shows 91% used at the point you displayed it above). Device swap (on disk) reflects how much of physical memory has been used (at some point, not necessarily now).

I would investigate your 4th server to see what can be done to easy memory pressure.
Possibly reducing kernel parameter dbc_max_pct or stopping unneeded processes.