SecureXL and CoreXL AVG Load

[HeikoAnkenbrand]

This oneliner (see spoiler) shows you the average CPU time of all CoreXL and SecureXL instances.

Click to Expand

SXLHACPU=`fw ctl affinity -l -a -v |grep Interface | awk '{print $6}' | sort |uniq | awk '{LINEHA = LINEHA "|CPU"$1} END  {print LINEHA }' | sed s/\|//`;SXLHACPU_C=`sar -P ALL 1 1 | grep Average: |grep -v CPU |grep -v all | awk '{ print "CPU"$2  " " 100-$8}' |egrep $SXLHACPU |wc -l`; SXLHACPU_S=`sar -P ALL 1 1 | grep Average: |grep -v CPU |grep -v all | awk '{ print "CPU"$2  " " 100-$8}' |egrep $SXLHACPU | awk '{sum+=$2} END {print sum}'`; SXLHACPU_LOAD=`echo "scale=2; $SXLHACPU_S/$SXLHACPU_C" |bc `; echo; echo "SecureXL AVG Load: $SXLHACPU_LOAD %"; U=`cpprod_util FwIsUsermode`; if [ $U == "0" ]; then CXL_LOAD=`top -b -n 1 |grep fw_worker | awk '{sum+=$9}END{print sum}'`; else CXL_LOAD=`top -b -n 1 |grep fwk0_dev|awk '{print $9}'`; fi;if [ $CXL_LOAD == "0" ]; then CXL_LOAD="1"; fi ;typeset -i CXL_C; CXL_C=`fw ctl multik stat | wc -l`;CXL_C=$CXL_C-2; CXL_LOAD_INT=`echo "scale=2; $CXL_LOAD/$CXL_C" |bc -l`; echo "CoreXL AVG Load  : $CXL_LOAD_INT %";

 Output:


Or use this tool for much more information on all your gateways :
Easy View Tool - View System Info for All Gateways Simultaneously

[Norbert_Bohusch]

Just tested in my lab with R80.40 und USFW and it is not working, as there is no fw_worker 😉

[HeikoAnkenbrand]

Hi @Norbert_Bohusch,

Now I check if the FW works as KMFW or as UMFW!
Accordingly I check the fw_worker or the fwk0_dev process.

THX
Heiko

[Norbert_Bohusch]

Thanks Heiko, but there is still a bug 🙂

Here the output (securexl avg always 100%):

Click to Expand

SecureXL AVG Load: 100.00 %
-bash: [: 6.2==: unary operator expected
CoreXL AVG Load : 2.06 %

I expect mpstat is different:

Click to Expand

[Expert@nb-ckp-gw1:0]# mpstat -P ALL
Linux 3.10.0-957.21.3cpx86_64 (nb-ckp-gw1) 04/14/20 _x86_64_ (4 CPU)

14:07:53 CPU %usr %nice %sys %iowait %irq %soft %steal %guest %gnice %idle
14:07:53 all 1.55 0.00 1.60 0.03 0.26 0.20 0.00 0.00 0.00 96.36
14:07:53 0 0.12 0.00 0.05 0.03 0.31 0.55 0.00 0.00 0.00 98.94
14:07:53 1 1.53 0.00 1.95 0.02 0.22 0.09 0.00 0.00 0.00 96.18
14:07:53 2 2.08 0.00 2.02 0.05 0.24 0.08 0.00 0.00 0.00 95.53
14:07:53 3 2.47 0.01 2.36 0.05 0.25 0.09 0.00 0.00 0.00 94.77

[HeikoAnkenbrand]

With the 3.10 kernel idle is on position 12 and with the 2.6 kernel on position 10.

I will modify the script again soon.

[HeikoAnkenbrand]

Hi @Norbert_Bohusch ,

Now with "sar" instead of "mpstat".
Should work now.

Can you test it?

Thanks
Heiko

[Norbert_Bohusch]

Tested it and also not working. But I have found the reason as well.
sar has the following syntax:
# sar [ options ] [ <interval> [ <count> ] ]
In kernel 3.10 R80.40 it is not using your "sar -P ALL 1" as count 1, but interval 1. But this should be correct according to the way the brackets are shown. In kernel 2.6 it is taken as count and therefor works.
You can fix it by using "sar -P ALL 1 1"

[HeikoAnkenbrand]

👍 Done!

Thanks
Heiko

[HeikoAnkenbrand]

I tested it:

✅ 3.10 /2.6 kernel 

✅ USFW and KMFW

✅ R80.10, R80.30

[Norbert_Bohusch]

Thanks, also working with R80.40 JHF 25 and USFW.