After the upgrade one of the Oneclick server has frequent (hourly) disconnects on the oneclick client and on the webapp that are lasting around 15sec.
The disconnects occur on the same time for the webapp and client and for every user using it.
This seems to have started since the upgrade to 21.2.12.
A network issue has been discarded.
The server has 4 cpu's and 32gb or ram, with 16GB ram for OneClick.
Only 5 users are using this Oneclick server, and there is no much more usage of CPU/RAM
No errors logged in dmesg or /var/log/messages.
Release : 21.2
Component : Spectrum OneClick
The OneClick server was running in a VM.
All kind of applications are sensitive to time changes or time difference between the VM and the host.
It is important to make sure the ESXI host and the VM sync its time against an NTP server.
Making sure that the ESXi host and the VM both synched its clock against an NTP server, resolved the problem.
The time difference was causing that every hour the VM time was synched with the HOST time, which had the clock 5 minutes ahead of current time.
This jump in time and then the correction made in the VM for the NTP client (or manually by the administrator) caused disconnections in OneClick that last for around 15 seconds.
Every Hypervisor vendor makes recommendations and provide best practices to avoid problems with the applications running in VMs due time differences with the HOST.