Hi,
We've been using HTTPAPI
successfully on our production AS400 for a few weeks, however
for the past week we've been having some issues.
Our prod box is on V7R3, and
actually had the latest PTFs applied to it on the Sunday just
prior to the issues occurring (the day before)...
I noticed that since this
upgrade we were getting "SetError() #49: SSL_protocol: Unknown
protocol 5000" appearing in the logs. Previously it was
showing as "Protocol Used: TLS Version 1.2". So
I saw another post in the forum where they'd been advised to
update HTTPAPI, and so I updated from V1.33 to V1.43 during
the week. After the HTTPAPI update we are now seeing "Protocol
Used: TLS Version 1.3" in the successful logs.
The following error is
occurring rather randomly throughout the day. We are running
our data posts jobs through a single data queue on the AS400,
and using YAJL to build each JSON data file that is then being
posted. Upon receiving this error, we basically delay for a
few seconds, rebuild the data file and try posting it again.
It typically works the second time.
________________________________________
HTTPAPI Ver 1.43 released 2021-04-19
NTLM Ver 1.4.0 released 2014-12-22
OS/400 Ver V7R3M0
http_setauth(): entered
http_persist_open(): entered
http_long_ParseURL(): entered
DNS resolver retrans: 2
DNS resolver retry : 2
DNS resolver options: x'00000136'
DNS default domain: xx.xxxxxxxx.xxx
DNS server found: x.x.x.x
DNS server found: x.x.x.x
Nagle's algorithm (TCP_NODELAY) disabled.
SNI hostname set to: xxx.xx.xxxxxxx.com
(GSKit) Peer not recognized or badly formatted message
received.
ssl_error(415): (GSKit) Peer not recognized or badly
formatted message received.
SetError() #30: SSL Handshake: (GSKit) Peer not
recognized or badly formatted message received.
_________________________________________
Following on from that, the
second issue is that we then find every 4 hours or so during
the day, we are receiving the following error;
The only way we've been able
to resolve this is to stop the data queue and restart it so
that it runs for the another 4 hours or so.
______________________________
HTTPAPI Ver 1.43 released 2021-04-19
NTLM Ver 1.4.0 released 2014-12-22
OS/400 Ver V7R3M0
http_setauth(): entered
http_persist_open(): entered
http_long_ParseURL(): entered
DNS resolver retrans: 2
DNS resolver retry : 2
DNS resolver options: x'00000136'
DNS default domain: xx.xxxxxxxx.xxx
DNS server found: x.x.x.x
DNS server found: x.x.x.x
SetError() #4: socket(): Too many open files for this
process.
______________________________
I know these are two separate
errors, but given they both started after the latest V7R3 PTFs
were applied, I suspect that the cause is related.
We will be needing to
increase the volume of data posts during the coming weeks, so
I'm a bit concerned that these errors will increase, and the
restarting of the data queue will also be required more
frequently during the day.
Your advice would be greatly
appreciated. Thank you!
Regards,
Corey.