CS1K Co-res linux working too slow with high CPU usage

[Juancho2015]

Hi there.

Several times I observed that CP-PM co-resident linux (6.0, and 7.0) server starts working slowly with high CPU usage, which results in high delays to get dial tone, IP phone registration, console and web access, etc.
Does anybody know the cause and possible solution?

Thanks in advance.

 

[KCFLHRC]

In my personal opinion. The CoRes option is junk. I have a few installed. I reload the software on them typically once a year. I would recommend checking the patch level. 6.0 is the worst in my opinion.

 

[KCFLHRC]

One other thing, the CPPM is not a very robust card and has a pretty slow processor. It has been replaced with the CPDC. However, I still believe the CoRes is a bad design. Cost effective? Yes. But not a great platform for heavy usage.

 

[Juancho2015]

Hi KCFLHRC,
I totally agree with you: Nortel screwed it up when built Linux based CS1K; VxWorks old versions worked far better and it was easier and faster to install them, and 6.0 was the worst of all.

The fact is that CoRes CPPM's usually work fine for small system but sometimes they seem to hang as described (with normal usage) and I was unable to find the reason nor the solution.

 

[KCFLHRC]

I have yet to find any rhyme or reason to it either. I have 2 CoRes 6.0 sites and you can pretty much bank on them needing the software reloaded once a year. I see the VAR partition getting full. Sometimes our TAC can clean it out through PDT but that just prolongs the inevitable.

 

[Juancho2015]

Alright, I´ll reinstall a 7.0 system in a new and larger HDD, then I´ll let you know.

 

[KCFLHRC]

I don't believe the HDD is going to help at all. It's the memory and processor that are lacking.

 

[Juancho2015]

I know what you mean, but RAM and CPU seem to be far enough 99.99999% of all working time.

This is a capture of TOP status screen:
top - 15:20:11 up 3 days, 21:21, 1 user, load average: 7.23, 8.41, 8.11
Tasks: 122 total, 1 running, 121 sleeping, 0 stopped, 0 zombie
Cpu(s): 16.9%us, 0.7%sy, 0.3%ni, 81.5%id, 0.0%wa, 0.3%hi, 0.3%si, 0.0%st
Mem: 2063312k total, 1809368k used, 253944k free, 255080k buffers
Swap: 2112536k total, 156344k used, 1956192k free, 720884k cached

This is a typical SAR log status:
12:00:01 AM CPU %user %nice %system %iowait %steal %idl
e
12:10:01 AM all 3.40 0.38 1.12 1.00 0.00 94.1
0
12:20:01 AM all 3.11 0.31 1.01 0.26 0.00 95.3
1
12:30:02 AM all 3.11 0.35 1.02 0.27 0.00 95.2
4
12:40:01 AM all 3.09 0.30 1.01 0.26 0.00 95.3
4
12:50:01 AM all 3.12 0.65 1.03 0.30 0.00 94.9
0
01:00:02 AM all 3.08 0.32 0.90 0.26 0.00 95.4
3