PA-5220 high SWAP memory usage

[agolob]

Hi,

We have from time to time higher SWAP memory usage on 5520 box. 

Not sure if this would be critical or if I can just higher the treshold values in monitoring system.

On reference box the swap memory is 3%, but on this one 85%.

Reference PA-5520
--------------------------------------------------------------------------------
Number of sessions supported: 4194302
Number of allocated sessions: 23251
Number of active TCP sessions: 5536
Number of active UDP sessions: 17425
Number of active ICMP sessions: 287
Number of active GTPc sessions: 0
Number of active GTPu sessions: 0
Number of pending GTPu sessions: 0
Number of active BCAST sessions: 0
Number of active MCAST sessions: 0
Number of active predict sessions: 6
Number of active SCTP sessions: 0
Number of active SCTP associations: 0
Session table utilization: 0%
Number of sessions created since bootup: 353923698
Packet rate: 5104/s
Throughput: 20602 kbps
New connection establish rate: 252 cps
--------------------------------------------------------------------------------

Problematic PA-5520
--------------------------------------------------------------------------------
Number of sessions supported: 4194302
Number of allocated sessions: 169554
Number of active TCP sessions: 70939
Number of active UDP sessions: 90414
Number of active ICMP sessions: 5642
Number of active GTPc sessions: 0
Number of active GTPu sessions: 0
Number of pending GTPu sessions: 0
Number of active BCAST sessions: 0
Number of active MCAST sessions: 0
Number of active predict sessions: 6369
Number of active SCTP sessions: 0
Number of active SCTP associations: 0
Session table utilization: 4%
Number of sessions created since bootup: 11960768196
Packet rate: 158076/s
Throughput: 994588 kbps
New connection establish rate: 5668 cps
--------------------------------------------------------------------------------

I mean there is more traffic, but not as much difference.

Any advise?

BR

[BPry]

'show system resources follow' will be the command that you want to run on the CLI to see what process is chewing through a higher than expected memory allocation. If SWAP is constantly high like that you likely have a malfunctioning process or some similar issue.

Out of curiosity have you attempted to restart the 5220 since you noticed the abnormal SWAP utilization? If one of the processes is malfunctioning for some reason that could clear the issue without spending a lot of time monitoring the memory utilization of the processes. 

[agolob]

To me this is not telling anything. "ikemgr" process is using the most of memory  - and this is only 5.3%.

I bet reload would solve this for a while, but it´s not a solution. Swap was slowly used more and more from the last upgrade on - so from the last reload. 

top - 13:43:06 up 63 days, 16:20, 2 users, load average: 0.59, 0.93, 1.02
Tasks: 246 total, 17 running, 227 sleeping, 0 stopped, 2 zombie
Cpu(s): 1.7%us, 1.4%sy, 0.2%ni, 96.4%id, 0.2%wa, 0.0%hi, 0.2%si, 0.0%st
Mem: 32640956k total, 32402084k used, 238872k free, 520652k buffers
Swap: 2007996k total, 1772292k used, 235704k free, 25208924k cached

PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
5786 root 20 0 1338m 228m 118m S 15.3 0.7 8944:52 useridd
5907 root 20 0 2678m 703m 10m S 5.3 2.2 3826:47 logrcvr
5905 root 20 0 2310m 1.7g 6396 S 2.7 5.3 908:07.72 ikemgr
871 root 20 0 0 0 0 Z 2.3 0.0 0:00.07 ping <defunct>
4840 root 0 -20 281m 16m 4720 S 2.3 0.1 148:22.13 masterd_apps
5927 root 17 -3 876m 69m 7056 S 2.3 0.2 1946:34 routed
4859 root 15 -5 95020 9696 3420 S 1.3 0.0 597:41.07 sysd
873 root 20 0 114m 9.9m 5008 S 1.0 0.0 0:00.03 masterd_batch
5921 root 20 0 480m 9328 4676 S 0.7 0.0 485:19.34 pan_dhcpd
20923 root 20 0 19496 1508 1092 S 0.7 0.0 0:08.06 genindex.sh
25342 nobody 20 0 393m 53m 7536 S 0.7 0.2 72:05.47 appweb3