Error in `/var/www/zoo-bin/zoo_loader.cgi': malloc(): memory corruption (fast)

[njunike]

In ADES v 2.0.21 we get the error "Error in `/var/www/zoo-bin/zoo_loader.cgi': malloc(): memory corruption (fast)" as soon as we try to deploy a new processor CWL file. For details see attached log file ades_log.txt. The URL to processor CWL file was cleaned up by me.

The issue does not occur in ADES v 20.0.17 and 20.0.19 but is systematically raised in 2.0.20 and 2.0.21.

[njunike]

ADES http response code is "501" with body

{
    "title": "NoApplicableCode",
    "type": "NoApplicableCode",
    "detail": "ZOO Kernel failed to process your request, receiving signal 6 = SIGABRT "
}

[gfenoy]

We tried reproducing the error, but we could not do so. We try here to determine what the software may be doing or guessing wrongly in this situation.

Can we have more details about this issue?

Is the CWL you are hiding the URL (which is perfectly fine) reachable from the environment where you are trying to deploy it without defining any proxy, security handling, or anything like this to get it fetched from this URL?

Did you modify anything from the default configuration to deploy ADES? If yes, what did you change there?

[njunike]

Hi Gérald,
thank you for looking into this! Yes, i can confirm that the URL to CWL file is public and reachable without any restrictions by ADES. In fact, same URL in ADES 2.0.19 works fine.

We have now published a demo processor in gitlab. With this demo processor in ADES 2.0.21 the same issue happens, see demo-processor.log. The CWL to this demo processors is pb-ieoto-demo-processor_app-package.cwl.
The CWL configuration of our ADES instance is in ADES_cwl_config.zip

[njunike]

To be clear: we do not change anything except the ADES version number (tag of docker image) when we switch between ADES 2.0.19 and 2.0.21. Do we need to adapt something else?

[rconway]

Note from @gfenoy...

I'm committed to finding a solution to the memory corruption issue.
While I couldn't reproduce the memory error with the example provided,
I did encounter a service execution failure during my investigation.
I'm continuing to investigate this issue and will update you once it works.