502 Bad Gateway - nginx

Discuss your pilot or production implementation with other Zimbra admins or our engineers.
Protack
Advanced member
Advanced member
Posts: 63
Joined: Fri Sep 12, 2014 11:34 pm

502 Bad Gateway - nginx

Postby Protack » Thu Jan 29, 2009 8:01 am

This issues has been plaguing us for sometime now. Has anyone started noticing this happen more often since 5.0.11?
God Bless,

Marty


Klug
Elite member
Elite member
Posts: 2358
Joined: Mon Dec 16, 2013 11:35 am
Contact:

502 Bad Gateway - nginx

Postby Klug » Thu Jan 29, 2009 8:07 am

Not at all.
What is your setup ?

How many zimbra-proxy, how many mailbox servers ?
Protack
Advanced member
Advanced member
Posts: 63
Joined: Fri Sep 12, 2014 11:34 pm

502 Bad Gateway - nginx

Postby Protack » Thu Jan 29, 2009 8:17 am

Nothing complex at the moment. We have a load balancer going into two proxys. Each of these proxys perform reverse for HTTP(s) and IMAP/POP. Then behind those we have only one mailstore at the moment. We have already eliminated the load balancer as the issue since we get this error while going directly to the proxys.
uxbod
Ambassador
Ambassador
Posts: 7811
Joined: Fri Sep 12, 2014 10:21 pm

502 Bad Gateway - nginx

Postby uxbod » Thu Jan 29, 2009 8:31 am

Anything reported in your Wiki :: Log Files when this occurs ?
Klug
Elite member
Elite member
Posts: 2358
Joined: Mon Dec 16, 2013 11:35 am
Contact:

502 Bad Gateway - nginx

Postby Klug » Thu Jan 29, 2009 9:27 am

Are DNS OK (split DNS if you're behind a firewall or even use the hosts file) ?
Did you check the mailstore configuration after upgrade ?

BonoboSLR had an issue while upgrading from
http://www.zimbra.com/forums/administrators/24260-solved-ngnix-help.html
Protack
Advanced member
Advanced member
Posts: 63
Joined: Fri Sep 12, 2014 11:34 pm

502 Bad Gateway - nginx

Postby Protack » Thu Jan 29, 2009 4:32 pm

We had this issue even with 5.0.11. I noticed that the version that Zimbra uses for NGINX is 0.5.30 and the latest stable version is 0.7.32. I wonder why Zimbra is so far behind. We are going to test in PD (pre-deployment) environment upgrading to 0.7.32 and see if this fixes the issue. Also DNS is properly setup. We do not NAT for our production environment either.
As far as errors in the log files there are the 502 errors in the nginx.access.log and then these in the nginx.log file:


2009/01/29 17:31:17 [warn] 20209#0: *114133 an upstream response is buffered to a temporary file /opt/zimbra/nginx-0.5.30-zimbra/proxy_temp/0/08/0000000080 while reading upstream, client: 192.168.36.200, server: zlm01.server.net, URL: "/service/zimlet/res/Zimlets-nodev_all.js.zgz?language=en", upstream: http://x.x.x.x:8080/service/zimlet/res/Zimlets-nodev_all.js.zgz?language=en", host: "mail.securemailhost.net", referrer: http://mail.securemailhost.net/zimbra/"

2009/01/29 17:31:18 [info] 20208#0: *114601 client x.x.x.x connected to 0.0.0.0:993

2009/01/29 17:31:18 [info] 20208#0: *114601 peer closed connection in SSL handshake while sending client greeting line, client: x.x.x.x, server: 0.0.0.0:993

2009/01/29 17:31:18 [info] 20208#0: *114602 client x.x.x.x connected to 0.0.0.0:110

2009/01/29 17:31:18 [info] 20209#0: *114416 client 10.88.0.17 closed keepalive connection (104: Connection reset by peer)

2009/01/29 17:31:18 [error] 20208#0: *114603 connect() failed (111: Connection refused) while reading response header from upstream, client: 24.101.117.84, server: zlm01.server.net, URL: "/service/soap", upstream: http://x.x.x.x:8080/service/soap", host: "mail.server.net"

2009/01/29 17:31:20 [info] 20208#0: *114609 client x.x.x.x connected to 0.0.0.0:995

2009/01/29 17:31:20 [info] 20208#0: *114609 peer closed connection in SSL handshake while sending client greeting line, client: x.x.x.x, server: 0.0.0.0:995

2009/01/29 17:31:20 [error] 20209#0: *114461 connect() failed (111: Connection refused) while reading response header from upstream, client: 192.168.36.200, server: zlm01.securemailhost.net, URL: "/zimbra?loginOp=logout", upstream: http://x.x.x.x:8080/zimbra?loginOp=logout", host: "mail.server.net", referrer: http://mail.server.net/zimbra/"
>

God Bless,

Marty
sn00p
Posts: 32
Joined: Fri Sep 12, 2014 11:47 pm

502 Bad Gateway - nginx

Postby sn00p » Thu Jan 29, 2009 11:41 pm

We encountered such error month ago in Firefox+Zimbra Web (5.0.10). Error has gone away after installing latest version of Firefox. It seems that it is browser+nginx communication error. Now it is difficult to say what Firefox version causes error. If you are interested I could try to search log files for warnings for that time.
Sysoev do a lot of work making nginx powerful and we are happy to help him.
VS_onD
Posts: 20
Joined: Fri Sep 12, 2014 11:44 pm

502 Bad Gateway - nginx

Postby VS_onD » Wed Feb 25, 2009 12:43 am

ok, I'm removing my post since my problems were probably due to an entry in hosts file...
so in order not to misguide anyone, I'm removing my questions.
sn00p
Posts: 32
Joined: Fri Sep 12, 2014 11:47 pm

502 Bad Gateway - nginx

Postby sn00p » Wed Feb 25, 2009 12:48 am

Nothing happens after deleting cookies. I've tried to repeat this bug recently but with no success :)

Nevertheless FF 3.0.6 on linux works fine now.
VS_onD
Posts: 20
Joined: Fri Sep 12, 2014 11:44 pm

502 Bad Gateway - nginx

Postby VS_onD » Wed Feb 25, 2009 12:58 am

Do you use self-installed certificates? ie ones that Firefox has to permanently or temporary accept?

Return to “Administrators”

Who is online

Users browsing this forum: Google [Bot], hudojnick, MSN [Bot] and 11 guests