DO280 Lab Environment reliability

[John_Daisley]

Does anyone have any tips for getting the lab environment working reliably in DO280? Seemingly every time I start the lab it takes 30-40 minutes before I can actually get a lab start script to run without errors. I often end up having to run the sync utility on the utitlity machine which 30 mins plus or rebuild the environment. It's quite frankly a rubbish learning experience as by the time I have the lab working I've wasted most of the time I have available for study. I've raised support tickets but their response is to run the sync utility but it seems a poor solution!

It's really affecting my exam prep and with this exam every minute of practice counts!

[Chetan_Tiwary_]

@John_Daisley The OCP cluster will take time to get its components ready before you can send requests to the API as we are using a Single-Node Openshift (SNO) setup and it take  a while for the VMs and all the OCP containers to startup.

Please do not run any lab start script until the cluster is up. Once it is up and ready - you can start with any lab exercise you want. 

You can set the auto stop timer as per your requirement so that the cluster will not shutdown and you can practice the labs till you want. Also adjust the auto destroy timer as per your need. 

You can start the labs well in advance before you want to sit and start for practising to compensate for the wait time. Remember that if your labs get stopped or destroyed - then is the wait time of approx 30 mins otherwise once the cluster is ready you can continue till it is stopped or destroyed. 

Please be extra cautious in ch03 where you work on authentication, if the cluster auth is messed up - many times it becomes easier to destroy the cluster and recreate ( ofcourse with the wait time ) than to spend hours in resolving/troubleshooting. 

[John_Daisley]

Thanks for the reply but it's not really a workable solution. I study as and when I have breaks from other things in my day and often have only have an hour here and there to dedicate to it. This prohibts starting the environment in advance as I don't know for sure when I will need it! 

Surely there must be a better way of doing this than having customers waiting like this! It completely ruins the learning experience and leads to frustration.

[Chetan_Tiwary_]

@John_Daisley I totally understand your situation, however the cluster wait time is there by design as of now. May be it will be improved in future course releases. I will forward the feedback to the course development team. Thanks for your understanding and patience!

[rhnoname]

Hi, 

I agree with @Chetan_Tiwary_ the first installation usually takes about 30 minutes, but once the OCP cluster has been installed and ready it should get up in 10 minutes or so.

From my experience with this lab, maybe when you edit the auth on 3rd chapter or when you create a custom project template and therefore the apiserver current pod is terminated and the new one starts, it could - in some cases - take more time than expected if you don't cleanup the applied configuration.

I hope it helps

[chennark]

The lab environment is highly un reliable off late. It is wasting ample amount time waiting for it stop/start and run into errors. Atleast 3-4 hrs has been wasted in the last couple of days. Is there certain times the lab infra is more reliable than other times?

[Chetan_Tiwary_]

@chennark Could you please elaborate more on the issue ? Did you verify if your OCP cluster was up or not before you start doing lab exercise ?

Please  share an error screenshot so that we can check and help you.

[John_Daisley]

I feel your pain @chennark 

I used to set a reminder half an hour before I intended to study to go and start the lab environment. That usually (but not always) meant it was ready for use by the time I wanted to study but it also meant I was burning allocated lab hours for no reason!

The real problem was when I wanted to do a bit of adhoc practice when I found myself with half an hour or so to spare. This was simply not an option because of the environment issues so in the end I set up a small cluster on AWS. I could start up and shut down that cluster on demand much faster than the lab environment would come up. Obviously there is a cost to running a cluster on AWS but in my opinion it's worth it to ease the frustration with the Red Hat system.

[chennark]

I have exam scheduled for early next week. Week end the time is the time to do most practice and not at all satified with lab infra.