Portworx in Azure crashes server nodes?


I’ve provisioned 3 Azure nodes with 4 cores each. I’ve provisioned 2 128 GB disks on each as outlined in your default deployment. Basically, I took all of the defaults. I have a bash script unzip a file and move some other files. In total, this only takes a few seconds to complete. 3 of 5 times testing this has resulted in the underlying node that was doing the disk activity to fail and restart. Azure support said that I need to deploy larger disks…that seems kind of absurd to me. Does Portworx not work on smaller disks, the size that you default to in your deployment? I can…just not unzip or copy files? Could this be true?

Also…as a side note, it doesn’t seem like the default monitoring works in the default deployment or there are many other steps that I need to complete. Prometheus and Grafana appear to be broken.

I realize that this is a free and unsupported product at this point. It seems like the basics should be functioning before I’d feel comfortable on moving forward with an larger Portworx engagement.

Also…the restricted licencing also means that I can’t configure a test environment to try things in if I have a production environment. That seems to make the product unusable for anything other than a demo. Is that intended? How could I possibly do any kind of Portworx upgrade in a test environment prior to running updates in production? Certainly you wouldn’t recommend that right? It seems as if there were some releases and upgrades in the past that were difficult for some.

Thank you…

Hi Tom, can you provide the PX log files from the crashed nodes? That is, the output of journalctl -lu portworx*

Also, the output of pxctl status.


I’ve been messaging Gou in private about providing these files. He hasn’t responded for a couple of days. I’m just looking for a way to provide them to your team. The log is about 13K lines. Not sure if there is sensitive data in the log.

Please email it to support@portworx.com and someone will follow up with you.

1 Like