Troubleshooting with tcp_client

[Euan]

So, I have a node from one of our CDOT 9.1 clusters that all of a sudden can no longer reach the mail relay to send autosupport emails. I've checked DNS, IP routes, traceroute etc - all good - nothing has been changed on the filer,  the partner node still sends emails fine.

"you'll need to try a telnet to port 25 mate - can't help otherwise" they say...

TELNET? if only we had something that easy!

Wait - tcp_client - that used to work in 7-mode...

Yes - set diag mode, ensure diag user is unlocked, connected to node shell and set diag mode here.

Issued command and got an bind error:

tcp_client 4 <source IP for e0M> 25 <mail relay IP address> 25 -c 1 -s 1024

bind failed: Can't assign requested address

I've tried every IP allocated to this node of the cluster - same error.

Does anyone have any suggestions as to what I'm doing wrong or an alternative to troubleshooting open port connectivity from a pesky filer node?

[Euan · Accepted solution]

What I should have done is spoken with my colleagues first who reminded me of the telnet command available through systemshell on each node;

remember to set diag mode and unlock the diag user.

CLUSTER::*> systemshell -node <NODENAME>

NODENAME% telnet <target ip> <port>

not sure why the tcp_client gives that error - but have the diagnostic I wanted.

Will leave this here so I can find it in another 12 months.