org.ldaptive.OperationException java.net.SocketException: Connection reset]; remaining name ''

[Klingenstein, Nate]

I’m sure a bad NIC would be, well, more bad.  Other tunnels to the box aren’t breaking.

I’ll check with the directory administrators about changing the timeout.  We are using a new load balancer endpoint for this.  It’s not a problem, but we had to back out the change last time because we didn’t have 3268 uniformly enabled(which I didn’t know about at the time and the escalation never reached me), which eventually exploded with dead connections taking the IdP with it, so management is skittish about anything involving WARN and LDAP with null strings.  I appreciate the sanity check.

From: users [mailto:users-bounces at shibboleth.net] On Behalf Of Daniel Fisher
Sent: Friday, April 14, 2017 2:04 PM
To: Shib Users <users at shibboleth.net>
Subject: Re: org.ldaptive.OperationException java.net.SocketException: Connection reset]; remaining name ''

On Fri, Apr 14, 2017 at 4:29 PM, Klingenstein, Nate <nklingenstein at calstate.edu<mailto:nklingenstein at calstate.edu>> wrote:
Caused by: java.net.SocketException: Connection reset
        at java.net.SocketInputStream.read(SocketInputStream.java:209)

Confirm whether or not that your directory (or load balancer) has a rule that drops idle connections after some period of time.
Or continue lowering the validation period until this warning goes away.
Of course you may actually have a bad NIC, but you'd see more than just LDAP connection problems.

--Daniel Fisher

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://shibboleth.net/pipermail/users/attachments/20170414/3f83ff85/attachment.html>