-
Notifications
You must be signed in to change notification settings - Fork 2
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
Deploy issue on local after first install #10
Comments
Hmmm, I'm kind of surprised that this is coming from the I have seen some Deploy's fail because not all the schemas had been written to disk and so when it attempts to restore one of the dependencies is missing that it's expecting not to. This can usually be fixed by reexporting schemas using the Deploy dashboard in the settings section, but I'm not entirely sure that's your issue here. For something to be null in the |
To make it more interesting. If we do a echo > deploy-export and after that do a echo > deploy we get a error. This shouldn't be possible since we shouldn't have any changes |
So if you deploy changes from dev locally, the first attempt will work, but the next time you restore locally, it will fail with this error? |
PS I'm assuming this is Vendr v2 and Umbraco v9 right? |
A little more information: It's Vendr 2.0.4 on Umbraco 8.17.1 When I cloned the solution from dev and started it for the first time it did a deploy-locally-initialize as expected. In between the 2 deploys I didn't touch anything. |
The same issue is happening on my collueges PC's by the way |
Ok, so yea, that is really weird. I'm really struggling to believe it's the EventBus that has the problem as before Does the project contain any sensitive information? as if not, I'm tempted to ask if I could be added to the cloud project to pull it locally and see if it does the same for me 🤔 |
What email address do you want me to add? |
The dev env is the environment with vendr installed |
Ok great, I've managed to replicate the issue so I'll have a dig into things and see what's going on |
That's greatish news. |
Ok, so I can see what is null, but I need to figure out why. I believe it might be because the initial restore is done in the web context, where as the other restores are happening in a background task. For some reason, a static variable we rely on being hooked up isn't in the background thread (which is weird, because it used to work this way before). I'll need to dig into this further tomorrow. |
Thanks Matt, it's always a great start to find out what's happening. |
So I've managed to slightly fix it, but not entirely. I was caching some items in Lazy properties and one of these was getting a null value and so it remembered the null value and errored on every deploy. But I have now at least fixed it so that it doesn't remember this and triggering a deploy from the back office does then succeed. But there is still another side to this problem that I'll investigate further. It appears that on a fresh boot (touch web.config) the Umbraco deploy is occurring before Vendr's I'll dig deeper tomorrow. UPDATE I've also pinged HQ to see if this is expected behaviour (Deploy triggering before IComponents have executed) |
So I think I'm going to need to wait for some input from HQ on this one as it does indeed look like Umbraco Deploy is triggering before the Vendr component is done registering. From what I can tell in DotPeek it looks like it's the Deploy component that triggers a deploy in it's initialize method which I thin runs before Vendr's due to it being registered first. I'll need to get some input from HQ as to how we work around that. I can look to release a nightly build of the current interim fix which should mean that even if the first attempt fails, subsequent attempts should succeed once the Vendr component had initialized. |
If we could have a temporary fix that would be a great solution. |
There should be a new 2.0.5-beta build on our nightly feed shortly 👍 https://nuget.outfield.digital/unstable/vendr/v3/index.json As mentioned above, deploy attempts after an app restart will still fail, but you can re-trigger a deploy from the back office which will succeed. |
I've raised an Umbraco deploy issue here umbraco/Umbraco.Deploy.Issues#91 |
Thanks Matt, One question for now. What package do I need to update? Only the core? Or the complete Vendr package? I assume this fix is in Vendr 2.0.5-beta0004? |
@FransdeJong sorry, it' the main Vendr package yes, so Vendr 2.0.5-beta004 as you mentioned👍 |
Looks like HQ agrees there is a problem with Deploy running too early. They are testing a PR as we speak so fingers crossed this should get resolved soon 🤞 |
I would just like to add, that I'm seeing this everytime a data extraction should run, so also when creating a new environment and even when pushing changes from local to cloud. |
@MichaelNielsenDK even when running the Vendr 2.0.5 beta mentioned above? |
PS I believe the extraction will fail for any deploy that causes and app-restart so I think that covers what you are seeing. The only ones that will succeed currently are ones triggered via the back office of an already running site. From what I can tell in the HQ issue tracker, a PR has been made and applied but I don't know when it is due for release. |
@mattbrailsford I haven't tried Vendr 2.0.5-beta004 And yes, my addition was just to make it clearer for anyone else, that you'll probably see this with any deployment. I can see that it should be fixed in the Umbraco Deploy upgrade coming out tomorrow |
Gotcha! Do let me know if the issue is still present after the Deploy update 👍 |
I've upgraded Umbraco.Deploy to 9.2.0 and Vendr to 2.0.5-beta0007, but I'm still having the same issue as described in #11. |
@MichaelNielsenDK is this only after an app pool recycle? Or do you get this error when triggering a deploy from the deploy dashboard in settings also? |
@mattbrailsford Running data extraction from backend was successful, so it would suggest that the issue is when the app pool recycles. |
Hmm, then it really make me think it's a similar kind of issue. I've commented back on the Umbraco Deploy issue with some more info that may be relevant to Andy so hopefully this might lead somewhere. |
Ok, so a recommendation in the HQ issue is for us to switch from using a It's not entirely clear to me whether this should just straight up fix the issue or if it's just part of the solution. If anyone wants to test this on their install though, it would be greatly appreciated. |
The issue I had is gone after installing 2.0.5-beta0010 👍 |
We have a very weird issue that we think we have pinpointed right now,
When we pull the solution from dev and run it it will deploy automatically and all is fine.
After the first deploy every deploy fails with this message:
The Uda looks like this:
On dev all is good and the files are exactly the same.
We tried to debug the error but it was hard to find the real issue since we don't know what happens in Vendr.Core
Please let me know how we can provide you with more information so we can find out where this issue originates.
The text was updated successfully, but these errors were encountered: