Join GitHub today
GitHub is home to over 40 million developers working together to host and review code, manage projects, and build software together.Sign up
Error in `Passenger core': malloc(): memory corruption #2089
Open source Passenger 5.1.2/Nginx, installed through Rubygems with RVM
While restarting Nginx, Passenger core seems to be reluctant to start. It gets stuck in a loop of attempting to restart the core. After maybe, an hour or so, the Passenger is up and stable. But as soon as I restart Nginx, the process repeats.
This started since today morning, apparently randomly. And restarting the server is not helping either.
I have attached the log, logged at level 7. Hope it helps figuring out my issue...
This is happening for us as well, across multiple different servers. Our stack trace implies it is related to the phone home for security updates as it fails establishing a connection to 188.8.131.52, which appears to be Passenger's security check service. Our log looks effectively the same as @HimanshuM's.
Edit: This started just today, after the release of 5.3.2.
This is likely due to an issue fixed in 5.1.6 where once enough security advisories had built up and the total length of the response from the security update check server exceeded ~500B then Passenger would crash. You can update to the latest version of Passenger to receive this fix, or disable the security update checker if you aren't going to use it.
edited to fix: version where this was addressed.
Can confirm we were bit by this as well running on
Agree with @sixfeetover that this issue should not be left closed unless the default setting is changed or docs can be updated to reflect that in order to leave security updated enabled you must be running
referenced this issue
Jun 13, 2018
Hi , I see the same bug.
I found the securitycheck URL on the SecurityUpdateChecker.h
And,check the URL.
It's not JSON response.
I can confirm this on Ruby 2.2, Passenger 5.1.4. Upgrade to Passenger 5.2.3 solved it.
It was a pretty weird issue since our servers that hadn't been touched for months started segfaulting one after another.
Man I'm glad to have found this issue because we were unable to figure out root cause and that was quite demoralizing.