If you goto https://docs.geoserver.org/latest/en/user/community/scripting/ you find a set of 2.19.x pages (rather than 2.21.x that is latest) that are not actually linked to anything (but they are still found in google) - is there any way to delete them? and I guess update the build process to make sure it builds into an empty directory.
Yes, we know - we cannot delete pages from S3 (we do not have credentials for that). The best we can do is upload placeholder pages; but we chose not to do so (as we were anticipating having control of geoserver.org domain.
Since that is not happening, I fully support adding placeholder pages to the docs marking content that has graduated, or retired.
Blah it is worse than that, all the doc jobs are broken:
geoserver-2.20.x-livedocs:
···
Geoserver-main-live-docs:
upload failed: target/main/en/api/1.0.0/coverages.yaml to s3://docs.geoserver.org/main/en/api/1.0.0/coverages.yaml An error occurred (RequestTimeTooSkewed) when calling the PutObject operation: The difference between the request time and the current time is too large.
Looks like it started happening on the 28th…
Once again we do have a new location for docs available at OSGeo, and we can change Jenkins publish location:
Does this error message represent a clock difference between servers? I wonder if recent European timezone change has caused Jenkins to be slightly ahead of itself…
···
–
Jody Garnett
Geoserver-main-live-docs:
upload failed: target/main/en/api/1.0.0/coverages.yaml to s3://docs.geoserver.org/main/en/api/1.0.0/coverages.yaml An error occurred (RequestTimeTooSkewed) when calling the PutObject operation: The difference between the request time and the current time is too large.
Looks like it started happening on the 28th…
Once again we do have a new location for docs available at OSGeo, and we can change Jenkins publish location:
Yeah, this is a time skew between servers. We ran into this issue somewhat recently (a month or two ago, I think?), and I put a fix in place that ought to have stopped the issue from coming up again (automatic job to refresh the time). Evidently, that didn’t quite work.
DST could have very well caused the issue, I’ll look into it when I next get a chance.
Does this error message represent a clock difference between servers? I wonder if recent European timezone change has caused Jenkins to be slightly ahead of itself…
Blah it is worse than that, all the doc jobs are broken:
geoserver-2.20.x-livedocs:
Completed 87.3 MiB/96.4 MiB (0 Bytes/s) with 420 file(s) remaining
upload failed: target/2.20.x/en/user/data/vector/properties.html to s3://docs.geoserver.org/2.20.x/en/user/data/vector/properties.html An error occurred (RequestTimeTooSkewed) when calling the PutObject operation: The difference between the request time and the current time is too large.
Geoserver-main-live-docs:
upload failed: target/main/en/api/1.0.0/coverages.yaml to s3://docs.geoserver.org/main/en/api/1.0.0/coverages.yaml An error occurred (RequestTimeTooSkewed) when calling the PutObject operation: The difference between the request time and the current time is too large.
Looks like it started happening on the 28th…
Once again we do have a new location for docs available at OSGeo, and we can change Jenkins publish location:
Yes, we know - we cannot delete pages from S3 (we do not have credentials for that). The best we can do is upload placeholder pages; but we chose not to do so (as we were anticipating having control of geoserver.org domain.
Since that is not happening, I fully support adding placeholder pages to the docs marking content that has graduated, or retired.
If you goto https://docs.geoserver.org/latest/en/user/community/scripting/ you find a set of 2.19.x pages (rather than 2.21.x that is latest) that are not actually linked to anything (but they are still found in google) - is there any way to delete them? and I guess update the build process to make sure it builds into an empty directory.
Juan was able to install the correct NTP for the version of ubunutu being used, and ask it to check against the cloud providers servers.
Looks like the documentation build are back.
Juan was able to install the correct NTP for the version of ubunutu being used, and ask it to check against the cloud providers servers.
Looks like the documentation build are back.
Yeah, this is a time skew between servers. We ran into this issue somewhat recently (a month or two ago, I think?), and I put a fix in place that ought to have stopped the issue from coming up again (automatic job to refresh the time). Evidently, that didn’t quite work.
DST could have very well caused the issue, I’ll look into it when I next get a chance.
Does this error message represent a clock difference between servers? I wonder if recent European timezone change has caused Jenkins to be slightly ahead of itself…
Blah it is worse than that, all the doc jobs are broken:
geoserver-2.20.x-livedocs:
Completed 87.3 MiB/96.4 MiB (0 Bytes/s) with 420 file(s) remaining
upload failed: target/2.20.x/en/user/data/vector/properties.html to s3://docs.geoserver.org/2.20.x/en/user/data/vector/properties.html An error occurred (RequestTimeTooSkewed) when calling the PutObject operation: The difference between the request time and the current time is too large.
Geoserver-main-live-docs:
upload failed: target/main/en/api/1.0.0/coverages.yaml to s3://docs.geoserver.org/main/en/api/1.0.0/coverages.yaml An error occurred (RequestTimeTooSkewed) when calling the PutObject operation: The difference between the request time and the current time is too large.
Looks like it started happening on the 28th…
Once again we do have a new location for docs available at OSGeo, and we can change Jenkins publish location:
Yes, we know - we cannot delete pages from S3 (we do not have credentials for that). The best we can do is upload placeholder pages; but we chose not to do so (as we were anticipating having control of geoserver.org domain.
Since that is not happening, I fully support adding placeholder pages to the docs marking content that has graduated, or retired.
If you goto https://docs.geoserver.org/latest/en/user/community/scripting/ you find a set of 2.19.x pages (rather than 2.21.x that is latest) that are not actually linked to anything (but they are still found in google) - is there any way to delete them? and I guess update the build process to make sure it builds into an empty directory.