The previous commit deployed correctly, the only difference is two numbers in the markdown.
One thing I did just before the problem started was that I modified some “Branch deploys” in the “Deploy contexts” area. Deleted some obsolete branches and added one new branch. But I didn’t touch any other setting, nor the master or stage branch which were already present in the “Branch deploy” list.
Any help is appreciated. This is a live site, which needs to be updated urgently.
thanks
Thanks for the suggestion @luke, it will surely be helpful in the future!
The local build has lead us to this error message that we haven’t been able to find a fix to: error Field "src" must not have a selection since type "String" has no subfields
The paths are correct, the naming of the images is also fine, the plugin order in gatsby-config.js is also correct (these were the suggestions from searching solutions for this error).
The thing I don’t understand, is why we get this error, when the site deployed fine with the same code (it’s currently live and running with the same code), nothing changed (only some markdown data, but nothing to do with images). But no successful deploy ever since, and I’d really need to push some new content.
OK, I just thought it had to be publicly available, but here goes:
I can see from the build log, that the node and the npm versions are the same for the successful and the failed builds, so I guess that shouldn’t be the issue.
Hey @luke and @fool, we have updated all gatsby related plugins in our repo and are completely losing it, for some reason cannot get gatsby to build correctly (both locally and on netlify) and therefore cannot deploy updates to our site at all. We need to put out some black friday content, and are way too late already .
This is the last deploy we tried:
It’s quite clear that the problem has to do with the gatsby-remark-relative-images plugin, because the errors are all connected to images that should be handled by gatsby-remark-relative-images.
The thing we don’t understand is how it is possible that we were able to deploy one day, and without making changes to any functional code, we fail to deploy ever since.
There was one amazing moment today, when we were able to run a successful build command locally, then deployed with the netlify-cli, but with the same settings, and the same code, we cannot repeat this miracle again.
Is there anything we can do to make it easier to find a solution to our situation?
After downgrading to gatsby version 2.17.1, the build process moves past the mentioned image error.
Thanks to this issue on the gatsby-remark-relative-image page.
We are now stuck with a graphql error and it has to do with how we handle languages: select is not a function
We seem to finally have a direction to move towards
@megant, we don’t have access to the repo (the build system does but our support team does not). This makes our troubleshooting of build errors limited to what we see in the logs.
(There can be exceptions to this. For example, if the error is not during the build but the pre/post processing stages of the build that Netlify controls. If there are errors in these build stages, we often do have additional details we can see in backend logs.)
In this case, I only see the same log error that you see. Does this same error happen locally?
Another thing to check is that we are using the same version of node and npm packages in our build as are being used locally. Are those the same as well? If not, we can research why not and get that corrected.
If there are other questions or more details to share, please send those anytime .
our problem persists, but with a twist.
I am developing on a Linux system and am unable to build nor deploy locally.
My colleague uses a Mac with the same node, npm, gatsby, gatsby-cli versions az I am using, and is able to build but not deploy locally.
These are the versions he is using:
node: 10.16.3
npm: 6.13.1
gatsby-cli: 2.8.14
gatsby: 2.18.5
Can the different operating systems have anything to do with the different result?
I set the node and npm versions in the netlify.toml file, so the build also runs with these versions.
We have not been able to push a single commit that netlify was able to build and deploy ever since I opened this thread.
My colleague has been able to deploy from a local netlify-cli. That is what saved us.
For some reason neither of us understand, netlify will not deploy successfully since the day I made the changes to our deploy context settings (maybe doesn’t have anything to do with our problem, but it’s the only thing that changed when deploys started failing).
Since then we have tried many versions of gatsby, npm and node, updated gastby plugins… downgraded plugins to versions used when netlify deployed successfully… no luck.
Our newest error is:
chmod: cannot access ‘/opt/buildhome/.gimme_cache’: No such file or directory
Lots of things can affect the result, but it is not expected that if you can’t build locally, we’d be able to
netlify-cli is “build on your system and send us the source code”, so no surprise that that would work well; it removes any chance for our build system to be involved and break. If the CLI is not working for you we can also troubleshoot that, but let’s pick what your goal is and work on that goal instead of working on multiple things in parallel. I think you’d rather build based on commits in our CI. Is that true?
We pushed a deploy from netlify-cli yesterday, the site was fine, haven’t done anything since, and now the site can’t be reached. We need very urgent help!
-EDIT: now turns out my colleague did try to fix our deploy problem since the last successful deploy. But any help would be appreciated from netlifys side, to get our site up and running ASAP.
It turns out that my colleague installed a new version of node, and that may have caused some problem, but even though he has downgraded to the original version, our problem persists.
I am unable to build, but my colleagues local build is successful. It is a question how this is possible, but it’s less important then the other problem we are facing.
My colleague is able to build and locally serve the site, but when he deploys from the netlify-cli, instead of the site, there is a “page cannot be found” netlify error page.
We currently re-deployed an earlier successful deploy on the netlify admin, but are unable to refresh the site, and deploy anything new.
Is there anything we can do to find the cause, and fix our issue? Do you need any more specific information from our side to get closer to a solution?
If you believe the issue is related to a change in the node version, you can make sure that our buildbot uses the same version that your colleague is using (which builds successfully) by setting the NODE_VERSION env var, as mentioned in this doc about managing dependencies. I recommend making sure you are able to build locally yourself. Then trying my suggestion. The next step after that would be to try building your site locally using the same build-image our buildbot uses.
.