haproxy weired error
Can anyone please advise on this one. I have haproxy running for a while. Had exchange before and now have iredmail. My iphone sync and works well but can`t get my android phone to see the same server. According to syslog, here is the msg :
Message: ip.addr.of.android:61105 [10/Jun/2018:22:10:38.434] HTTPS-merged Mail_https_ipvANY/mail 1/1/2 0 SD 2/2/2/2/0 0/0
Can anyone advise please as I can`t work this one out
Well.. haproxy manual says this for your SD state..:
SD The connection to the server died with an error during the data transfer. This usually means that haproxy has received an RST from the server or an ICMP message from an intermediate equipment while exchanging data with the server. This can be caused by a server crash or by a network issue on an intermediate equipment.
As for the 2 millisecond total connection time and 0 byte reply from the server.. it doesn't look like much happened there..
I think your haproxy does not do anything with the traffic like offloading/re-encrypting right? Just plain 'tcp' through frontend>backend>server
What do the iredmail logs say.?
When not going through haproxy the phone can connect to the iredmail directly without issue?
There is no error on Iredmail logs hence I am going crazy. I know that haproxy is stable and working as the iphone goes through fine. If I go on the lan and then connect the android phone, it works so most likely the issue is something between the way the android phone is trying to get the data :-(
The proxy is set to plain tcp through. Would you like me to pm you the config just in case you can spot an error. Iredmail uses SOGo for eas for the phones connection.
If you can PM me the config thats okay ill take a look at it. (not sure yet how PM works in this new forum..)
As for diagnosing the issue, perhaps try and tcpdump the traffic, on both wan and lan sides of haproxy and see if any 'data' passed in is also being send out on the other side? (check tcp data lengths, not just the 'packet count', as that could validly be different..)
@PiBa Did you ever resolve the SD errors?
i dont recall the specifics about rajbps his issue.. its been over a year! , but afaik there is no issue with haproxy itself..
@PiBa Thank you for your response. I was asking because I am experiencing a similar issue using PostgreSQL 11 on the back end. Every time we close the session, HAProxy logs an error with a termination state of SD. I'm just curious if it's a configuration issue with either HAProxy or PostgreSQL. If it's a harmless error, then is there any way to suppress the error in the log?