kubectl command respond randomly slow in AIO- Duplex setup of STX on VM uisng virt-manger

[Yatindra Shashi]

Brief Description
-----------------
In StarlingX 5.0 duplex setup which is running on the VM using virt-manager on the ubuntu machine, has issue on response of Kubectl command randomly slow.

Severity
--------
Major

Steps to Reproduce
------------------
1. Install STX AIO-DX in the VMs using virt-manager
2. Storage is SATA drive
3. On every first login after automatic logout, kubectl command takes longer time to respond. Sometimes about 45 sec.
4. Even after certain time of not use again kubectl response slow

- This even happens when i don't use controller-1(controller-1 is in poweroff state)
Expected Behavior
------------------
Kubectl should response fast

Actual Behavior
----------------
- Kubectl response slow randomly

Reproducibility
---------------
In my VM setup it happens everytime

System Configuration
--------------------
2 Node : stx 5.0
AIO-DX

Branch/Pull Time/Commit
-----------------------
stx 5.0

Timestamp/Logs
--------------
logs attached
- kubernetes pods seems to be restarted multiple time like

[Yatindra Shashi (yshashi)]

Screenshot of image of kubectl

[Yatindra Shashi (yshashi)]

tar file for controller-0

[Yatindra Shashi (yshashi)]

I keep on getting error like ping Ack does not arrive on time as attached screenshot for drbd. Is it because of e1000 mgmt network interface being slow compare to virtio?

[Yatindra Shashi (yshashi)]

From my debugging, I think its because of Hard drive. I have hdd drive with 5400 rpm, I will check with ssd and update if it fix the issue.

[Ghada Khalil (gkhalil)]

screening: marking as low priority due to lack of activity