Sync Between Gateway & Portal
Article ID: 270257
Updated On:
Products
Issue/Introduction
All the Scheduled tasks between Portal & Gateway runs only one node from each cluster. What I wanted to know is what is the impact if those are selected to run on all the nodes instead of one ?
Also will it change the Cpu & memory usage ?
Environment
Release : 10.1
Resolution
This is a change we would NOT recommend.
Options for scheduled tasks are One or All.
Using All would mean that every node in a cluster will be attempting to perform exactly the same task at ~exactly the same time. This shouldn’t result in any data integrity issues (duplicates) but it does have potential performance issues.
The cluster wide properties and the data is shared by all the gateways in the cluster, the CWP property portal.applicationincrement.start is the epoch time of last sync
ALSO
The Gateway that runs the schedule task would be the cluster master. You can check this and get the nodeID in the SSG DB then find the gateway from the cluster_info table:
mysql> select * from cluster_master\G;
*************************** 1. row ***************************
nodeid: 7f7533ea2a54459c9b7af7206b338663
touched_time: 1689970719455
version: 6193344
1 row in set (0.00 sec)
ysql> select * from cluster_info\G;
*************************** 1. row ***************************
nodeid: 034a660c114d4f40bba527b9a0d32b84
mac: 00:50:56:A0:4D:E4
name: Gateway1
address: <GW_IP>
esm_address: <GW_IP>
multicast_address: <MA_IP>
uptime: 1689954498702
avgload: 0.18
statustimestamp: 1689970929166
*************************** 2. row ***************************
nodeid: 7f7533ea2a54459c9b7af7206b338663
mac: 00:50:56:A0:65:36
name: Gateway2
address: <GW_IP>
esm_address: <GW_IP>
multicast_address: NULL
uptime: 1683661008790
avgload: 0.11
statustimestamp: 1689970928245
2 rows in set (0.00 sec)
Feedback
