New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Terraform crash due to govmomi property collector issue on dvportgroups #952
Comments
FYI, i found that vCenter responds with a bad XML, which causes the panic. More details here, on the govmomi issue. Still hunting why exactly does vCenter respond badly and then stop responding (without saying anything in the logs), and why is this happening since moving to terraform 0.12 and terraform-provider-vsphere 1.14/1.15. |
Hi @sofixa, Unfortunately, as you noticed yourself, even if we limit the properties we ask the Collector to return you'd still get the same issue. Until this is resolved upstream we can't do much. I was looking at the govmomi issue you linked and looked at the logs you attached and I noticed this bit:
Is there a change you have some kind of firewall or proxy between the host running terraform and vsphere ? Could it be possible that the connection gets dropped because of its size and we only get back half of the response? If you do have something in place can you try reproducing the issue by running terraform from a host in the same network as vsphere? |
Hi @koikonom , thanks for the response. IMHO the I tried running terraform from a machine on the same vlan as both vCenters, outside of docker as well, just in case. I've also updated the vCenters to the latest 6.7 version, still no dice :( |
I'm going to close this one out since its an upstream issue. Once it is fixed in govmomi, we'll pull in the new version. Thanks @sofixa for keeping on getting the info to the govmomi team! |
I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues. If you feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. If you feel I made an error 🤖 🙉 , please reach out to my human friends 👉 hashibot-feedback@hashicorp.com. Thanks! |
Terraform Version
Affected Resource(s)
Please list the resources as a list, for example:
vsphere_distributed_port_group
(i think)Terraform Configuration Files
Debug Output
The debug output is 796655 lines so i'll only provide the relevant part:
Expected Behavior
After an upgrade to 0.12 and a refactor of the modules to pass most shared data sources as parameters instead of having them in each module (to save time and resources), i'm doing a plan
Actual Behavior
which fails on the
refresh
stage (even a simple refresh results in the same). It's always on refreshing dvportgroups, random ones, and after the crash the rest of the refreshes fail withrpc error: code = Canceled desc = context canceled
probably because the provider-vsphere died.This is a project working perfectly (albeit slowly on 0.11) with slight refactoring (but nothing that should be seen during the
refresh
)Steps to Reproduce
Not 100% sure, but basically:
Potentially the fact they were created by 0.11 on an older provider version could change something, not sure.
Important Factoids
It's a huge project, and we have a similarly created dvportgroups on other, smaller projects, and they work fine (but they were created with 0.12, and are on a completely different group of vCenters, with slightly different versions, so maybe that changes something).
Since it seems like a basic govmomi issue, similar to the ones patched by #840 , i tried applying the same logic by limiting the returned properties by the properties collector, but that changes nothing (probably because those are the only two properties available for the type):
References
Basically the same thing, but for network_interface:
Thanks,
Adrian
The text was updated successfully, but these errors were encountered: