my last measurements at the time were giving a 4-5 times difference in the metadata rate by enabling it for the internal control channels.

This makes very little sense to me unless you are somehow disabling HTTP keepalive (and hence incurring a few extra roundtrip overheads per metadata operation when HTTPS is used). If that's true, the more reasonable approach would be to fix to make HTTP keepalive work - in that case, you would also gain speed because you avoid the TCP handshake as well.

The feature was untouched.

Agreed - there's nothing to do with the TLS layer here. This bug has existed since early on in the 4.x series (whenever we first added macaroon support I suspect ... the second code path is probably 6 months old as we improved EOS support). No one noticed it because, outside @riccardodimaria, no one has tried using both tokens and this secretkey feature.

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

[ { "@context": "http://schema.org", "@type": "EmailMessage", "potentialAction": { "@type": "ViewAction", "target": "https://github.com/xrootd/xrootd/issues/1251#issuecomment-658828106", "url": "https://github.com/xrootd/xrootd/issues/1251#issuecomment-658828106", "name": "View Issue" }, "description": "View this Issue on GitHub", "publisher": { "@type": "Organization", "name": "GitHub", "url": "https://github.com" } } ]

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