-
Notifications
You must be signed in to change notification settings - Fork 536
KeyError sh256 #5093
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
Comments
Is this problem fixed? |
Also encountering this, any update on a fix? |
Is there a work around for this problem? |
Is this problem fixed? |
Did some investigation on this: We are running cve-bin-tool based on official python:3.11 docker image. Builds are failing since the beginning of this week. I compared used pip dependencies between our last successful and the failing builds. No differences found. Switching to latest python:3 image did not solve the issue. I guess the problem arises from latest definition updates? |
We started to use the following as a temporary workaround.
|
I have the seen the same problem (running Python 3.10). A bit of digging has shown that some of the files have missing metadata
However, after modifying the code to add a check that the sha256 field was present, I still encountered an error with the data for 2022. It looks like the meta data has a SHA value but as there is no file avaialble, so there is a SHA mismatch
Looking at the mirror site you can see two years have 0 length gz files. Adding a check to detect a zeo length file, prevents the SHA mismatch and allows the database update to proceed. |
Workaround (removed sha256 check). nvd_source.py after before after
|
This is not fixed. It's going to be at least a few more hours before we can get the mirror updated. Current workaround is to use In theory changing data sources is safer than turning off signature verification, but the signatures aren't really the greatest validation at the best of times so I wouldn't blame folk for using that workaround. |
pinging @warthog9 so he can see the details before he goes to mess with the mirror |
Ok doing some poking and digging, looks like something weird is going on upstream of me, which is not exactly helpful. When I'm mirroring out the files from NVD itself I'm apparently getting 0 length files on occasion now. LOOKS like if I mirror it, then mirror it again, I get things sorted out and no longer get 0 length files. Not ideal, and it makes me wonder what's changed somewhat recently to be getting those kinds of issues. Note: Should be fixed on my end, I'm running a couple of checks now though |
Thank you all! |
In case anyone's wondering: our cache in github did not update correctly. I'm re-running it in case it was just a mirror propagation issue, but we may still have a problem. |
Looks like we hit a problem with the update code over the weekend, so the github cache is currently broken. I'm re-running it to see if it's already resolved (last time it was a data issue) but I'll be back to take a look in an hour or so.
Current workaround
Current workaround is to use
-n api2
and set annvd_api_key
, which allows you to get the data from nvd directly via their API. If you've already got most of a cache this should be great, if you're starting with no data you may discover that this is very slow (so if you're changing a job that has a timeout set, you may need to lengthen the timeout)The text was updated successfully, but these errors were encountered: