Hello,
I'll comment from the FTS side.
Going by @abh3 latest statement, I'm not yet convinced:
That this is a Davix problem and Davix only works when DAVIX_DISABLE_SESSION_CACHING=true and it does not matter what the server's setting is in that case.
Davix session caching is enabled by default. To disable it, one has to set the DAVIX_DISABLE_SESSION_CACHING
variable. The value does not matter (davix/neonsessionfactory.cpp):
static bool isSessionCachingDisabled(){
return ( getenv("DAVIX_DISABLE_SESSION_CACHING") != NULL);
}
Leaving outside on why Davix only checks if the variable is set or not (and not the actual value), this means the previous table now turns into:
client | server | works? |
---|---|---|
DAVIX_DISABLE_SESSION_CACHING set | http.tlsreuse missing | no |
DAVIX_DISABLE_SESSION_CACHING set | http.tlsreuse missing | yes |
DAVIX_DISABLE_SESSION_CACHING set | http.tlsreuse on | no |
DAVIX_DISABLE_SESSION_CACHING set | http.tlsreuse on | yes |
DAVIX_DISABLE_SESSION_CACHING set | http.tlsreuse off | yes |
DAVIX_DISABLE_SESSION_CACHING unset | http.tlsreuse off | no |
To me, this looks like noise (previously, ~38% failing transfers failing, were mentioned).
If the session caching had a definitive effect, we should've seen it.
—
Reply to this email directly, view it on GitHub, or unsubscribe.
You are receiving this because you commented.
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