Details
-
Bug
-
Status: Closed (View Workflow)
-
P2
-
Resolution: Done
-
None
-
-
Folijet Sprint 114, Folijet Sprint 115, Folijet Sprint 116
-
3
-
Folijet
-
R1 2021 Hot FIx #2
-
Yes
Description
Seen in vagrant box folio/testing - version 5.0.0-20210412.6124 - but also for earlier releases.
Module versions: mod-source-record-manager-3.1.0-SNAPSHOT.388 , mod-inventory-16.4.0-SNAPSHOT.358 .
In my test I am using vb.cpus=12 and vb.memory=18000. No swapping happening.
The system at boot deploys all modules.. This takes a couple of minutes. After this point, no user is logged on. However, the CPU load is 2 (200%) in total for the system. What is measured here is "TIME" .. Ie the total CPU usage user+system for a process.
After 10 minutes, the system is surely turned to "idle" state and the top 4 contenders are: mod-source-record-manager 3:44 (3 minutes 44 seconds), mod-inventory 3:39, kafka 2:07, mod-pubsub-server 0:57. .. further down is list is Okapi 0:17.
After one hour the stats are as follows: mod-source-record-manager 24:22, mod-inventory 24:01, Kafka 11:41, docker-proxy 5:50, mod-pubsub-server 3:25. Down the list is okapi with 0:37.
Would expected that mod-inventory, mod-source-record-manager would not use any significant CPU after 10 minutes. Okapi is running mod-circulation jobs every minute.. And has done work (20 seconds ouf of 3000 seconds in total ~ 0.03% ).
mod-source-record-manafer OTOH has used abt 20 minutes of CPU of in 50 minutes (40%).
System load like this is not directly seen by users, but it will slow the system down as a whole.. Not to mention, add costs to hosting.
There is no log in okapi.log to indicate what these modules - mod-source-record-manager and mod-inventory would be doing in a system with no importing, exporting or user activity.
The following command line was executed to gather information:
while true; do uptime >>stat2 ps -w -a -A -o pid,user,time,args --sort time >>stat2 sleep 60 done
Attached is stat2 for 2 hours of operation.
It's not clear whether is an issue in mod-inventory or mod-source-record-manager.. One of these modules initiates a message quite often (or something).
How it looks like on top:
top - 11:11:13 up 2:14, 1 user, load average: 1.70, 1.97, 1.99 Tasks: 468 total, 1 running, 467 sleeping, 0 stopped, 0 zombie %Cpu(s): 6.0 us, 4.0 sy, 0.0 ni, 89.1 id, 0.1 wa, 0.0 hi, 0.8 si, 0.0 st MiB Mem : 17578.1 total, 3592.0 free, 11755.8 used, 2230.3 buff/cache MiB Swap: 980.0 total, 980.0 free, 0.0 used. 5495.0 avail Mem PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND 14864 vagrant 20 0 2314012 251620 16268 S 45.5 1.4 55:00.09 java 7255 vagrant 20 0 2309856 194572 15612 S 44.9 1.1 54:16.52 java 15785 root 20 0 6438640 640892 11768 S 25.7 3.6 25:46.61 java 15396 root 20 0 1216856 13932 2812 S 12.9 0.1 13:20.60 docker-proxy 12130 vagrant 20 0 1945208 277152 17484 S 5.3 1.5 7:04.01 java 13077 vagrant 20 0 1798936 233004 16380 S 3.6 1.3 5:20.91 java
TestRail: Results
Attachments
Issue Links
- blocks
-
MODSOURMAN-488 Release v3.0.8 (R1 Iris Hot Fix 2)
-
- Closed
-
- defines
-
UXPROD-3023 NFR: R2 2021 Juniper Data Import Stabilization and Reliability work
-
- Closed
-
- relates to
-
MODINV-423 Excessive CPU usage in a system with no user activity
-
- Closed
-