Forum Discussion
Josh_41258
Nimbostratus
NimbostratusBasic SMTP Virtual Server
I am trying to configure a basic virtual server to load balance a few SMTP servers. I created a new virtual server listening on TCP/25, no profiles, no OneConnect, AutoSnat, and other default settings. When attempting to connect to the virtual server, the connection just hangs for a few seconds and then disconnects. Upon applying the standard "smtp" SMTP profile to the virtual server, it is immediately disconnected. I do not see any connection attempts on the individual SMTP servers. It's like the F5 is not passing the connection onto the pool.
Can someone give me any suggestions about why this is happening/how to fix/troubleshoot?
Thanks,
Josh
- This smells a bit like a routing issue of some sort, though I'd start with the usual suspects:
1. Can the LTM see the SMTP servers? What happens if you log into the LTM and telnet to port 25 on the SMTP servers?
2. Do the SMTP servers show as 'up' in their pool? If you're using a monitor, is it correctly (or incorrectly) marking them up or down as appropriate?
3. What does a tcpdump show when a connection is made on the VIP/external side? and the internal side?
4. Anything oddball in the logs on your SMTP server or in /var/log/ltm on the LTM?
One or more of these should get you on the right track.
The 'smtp' profile comes into use if you're running ASM or ASM - it's used to hook the security features of those 2 products, and is otherwise not needed. It works the same way as an HTTP class profile does, except for SMTP. 
Josh_41258I have several other virtual servers that are on the same subnet that the SMTP servers are on with no problems. The internal interface of the F5 is on this same subnet. Here are the answers to your questions:
1. Yes, I can successfully telnet to the SMTP server from the F5.
2. Yes, when the SMTP health monitors are enabled, the servers are showing as "up" I have also tried disabling the health monitors with no success.
3. Nothing to capture on the internal side.. This is what I see on the external: .. SMTP server is 172.26.101.48, source is 172.26.101.56, VIP is 172.26.137.134
No. Time Source Destination Protocol Info
1 0.000000 172.26.101.56 172.26.137.134 TCP 47127 > smtp [SYN] Seq=0 Win=5840 Len=0 MSS=1460 TSV=3704817064 TSER=0 WS=7
2 0.000011 172.26.137.134 172.26.101.56 TCP smtp > 47127 [SYN, ACK] Seq=0 Ack=1 Win=4380 Len=0 MSS=1460 WS=0 TSV=3253379147 TSER=3704817064
3 0.000285 172.26.101.56 172.26.137.134 TCP 47127 > smtp [ACK] Seq=1 Ack=1 Win=5888 Len=0 TSV=3704817064 TSER=3253379147
4 12.600101 172.26.137.134 172.26.101.56 TCP smtp > 47127 [RST, ACK] Seq=1 Ack=1 Win=4380 Len=0
4. Nothing in /var/log/ltm about this specific problem, and nothing in the individual SMTP server logs either. I don't believe they are seeing the connection attempt.
Thanks!- Josh_41258OK.. odd, but I deleted the virtual server, recreated it with the same options, and it works now.
Josh - That is weird. Once in a blue moon, I've run into similar little deals. Never have been sure why deleting and re-adding has fixed them.
- dennypayne
Employee
That fix has worked for me on occasion as well, I suspect it has something to do with an out of order line (or section) in bigip.conf that for some reason doesn't prevent the config from loading, but doesn't actually work either. I haven't been able to diagnose what causes that.
Denny 
Thiago_FigueiroI was having the same problem and came across your message.
Re-created the virtual server and now it works.
I had been fiddling with a NAT setting for the same VIP so it might have been that.
Thanks for sharing.
