High Management CPU

[SOC_CSG]

Hi

Since upgrading a PA-2050 to 5.0.8 I noticed that the management CPU is always above 70% and DataPlane is under 5%. Do this behaviour is normal? Management CPU sometimes goes to 99%.

Best regards

Gonzalo Arroyo

[kadak]

Hello Gonzalo,

Its ok to see management CPU spike - but only during certain times like commits, generating tech supports, report generation and similar processes. BUT, if it stays to 70% for the most part of the day then we would need to check what process on the management plane are causing the management CPU to go high.

Lets see how the management plane looks resource-wise.Could you give us an output for the following commands:

> show system resources | match srvr

> show system resources | match total

> show system disk-space

Has anything changed since the upgrade?  How did the CPU look prior to the upgrade (in other words expected MP CPU usage)?

Thanks and regards,

Kunal Adak

[VinceM]

Hi,

Sometime this issue happen when database is running out of memory. try to reset your log.

In parallel restart you rmanegemnt server with: debug software restart management-server

Hope help.

v.

[SOC_CSG]

Hi

I have restarted the management-server but doesn't solve nothing.

The output of the commands requested are these:

telindus@fw1orgt(active)> show system resources | match srvr

7438       20   0  386m 220m 5896 S    2 22.7   3:23.14 mgmtsrvr

2450       20   0 234m  46m 7968 S    0  4.7 118:59.52 devsrvr

telindus@fw1orgt(active)> show system resources | match total

Tasks: 102 total,   2 running, 100 sleeping,   0 stopped,   0 zombie

Mem:    995872k total,   942036k used,    53836k free,    22880k buffers

Swap:  2212876k total,   311428k used,  1901448k free,   205648k cached

telindus@fw1orgt(active)> show system disk-space

Filesystem Size  Used Avail Use% Mounted on

/dev/sda2 3.8G  1.4G  2.2G  39% /

/dev/sda5 7.6G  4.1G  3.2G  57% /opt/pancfg

/dev/sda6 3.8G  2.6G  1.1G  71% /opt/panrepo

tmpfs 487M 67M  420M  14% /dev/shm

/dev/sda8 125G   73G   46G  62% /opt/panlogs

Thanks in advanced

[mikand]

Do you have anything that is polling the mgmt plane, or is running netflow, crl-checks (ssl), userid or such?