When a redirector is in the loop, it fails when redirected to a 5.x data server but works when redirected to a <= 4.12.3 server. Is that correct?

Exactly, that's the only way I can reproduce it. Of course, it does not mean that this is the cause — it could be that this triggers the correct "timing" / delays between reusing cached sessions to cause this.

It would seem you would get the same problem with a 4.12.4-RCx data server which uses the same code to do this. Perhaps you can try that to verify that it's a problem in this new code.

I will try. I can not promise a timeline yet, though: The debug information was collected in our production setup after an upgrade with fingers crossed, since I was unable to see this issue with our test setup, since our test setup currently is a single node only. So I will first have to expand our test setup to two nodes to freely test also RCs there, I did not anticipate there would be issues only visible with a redirector, but of course issues always hide in the single difference you have between test and prod ;-).

Unfortunately, the newly added code does not propagate the debug settings to the underlying x509 library. You can force this by configuring gsi in xrootd with the proper debug flags set there.

You mean the -d setting here?
https://xrootd.slac.stanford.edu/doc/dev50/sec_config.htm#_Toc42289901
I'll set these after reproducing this in our test environment (will take at least two days though to get a new Grid certificate for the node).

—
You are receiving this because you are subscribed to this thread.
Reply to this email directly, view it on GitHub, or unsubscribe.

Use REPLY-ALL to reply to list

To unsubscribe from the XROOTD-DEV list, click the following link:
https://listserv.slac.stanford.edu/cgi-bin/wa?SUBED1=XROOTD-DEV&A=1