Performance guest directory
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
zVM Cloud Connector |
New
|
Undecided
|
Unassigned |
Bug Description
We have around 60 guests deployed through Cloud Connector. We want to get data of these guest sometimes in a report. For these we made a Python script for collecting. This script however runs couple minutes. We call /guests/SLD510T, this costs around 2 sec.
I tried the smcli version (times):
time smcli Image_Query_DM -T SLD510T
USER SLD510T LBYONLY 4096M 4G G
INCLUDE LINUX
COMMAND DEF STOR RESERVED 0M
CPU 00 BASE
CPU 01
IPL CMS PARM AUTOCR
MACHINE ESA 10
DEDICATE 5C50 E014
DEDICATE 5C51 F014
MDISK 0100 FB-512 126829152 25165824 LNXOS2
MDISK 0403 FB-512 30165856 104857600 LNXOS3
MDISK 1300 FB-512 126829152 1000000 LNXOS4
*DVHOPT LNK0 LOG1 RCM1 SMS0 NPW1 LNGAMENG PWC20190403 CRCXX
real 0m2.060s
user 0m0.001s
sys 0m0.002s
So this give indeed a timing of 2 seconds.
Is this normal because a call with DIRMAINT on z/VM to get this data is under 0.1 s?
Or is this the interface penalty with persmapi or zthin?
Or could this be a failure in the setup of our Cloud Connector?
Thanks
Herald ten Dam
I did some test locally ,it takes 1 sec, from my expereince, it's unlikely be caused by zthin for long delay, it might be caused by zthin socket connect to SMAPI and lead to some network latency
and 0.1 sec you mentioned might be DIRMAINT only? as the time will be smcli -> SMAPI -> DIRMAINT (actually DIRMAINT is more slow, so SMAPI has an internal cache and return very fast)
IDENTITY OPNCLOUD XXXXXXXXX 8G 8G G
BUILD ON OPNSTK1 USING SUBCONFIG OPNCLO-1
BUILD ON OPNSTK2 USING SUBCONFIG OPNCLO-2
CPU 00 BASE
CPU 01
CPU 02
CPU 03
MACHINE ESA 4
IPL 190 PARM AUTOCR
IUCV ANY MSGLIMIT 512
OPTION LNKNOPAS
CONSOLE 0009 3215 T
SPOOL 00C 2540 READER *
SPOOL 00D 2540 PUNCH A
SPOOL 00E 1403 A
* END IDENTITY OPNCLOUD
*DVHOPT LNK0 LOG1 RCM1 SMS0 NPW1 LNGAMENG PWC20161220 CRC▒"
real 0m1.032s
user 0m0.001s
sys 0m0.001s
yes, it take some time and it do need some