Details
-
Bug
-
Resolution: Done
-
Critical
-
None
-
None
-
Documentation (Ref Guide, User Guide, etc.)
-
The workaround for this issue is to configure the "HttpProtocolOptions" option set to "Unsafe" in httpd.conf
Description
This is a regression as 2.4.23-GA works.
Reproducer:
1. Set up httpd mod_cluster balancer with Ipv6 address with advertisement on.
2. Set up worker node (tested with EAP)
3. Checks on console if they correctly connect
Issue:
Httpd log:
[Wed May 10 04:41:31.930288 2017] [core:debug] [pid 15638] vhost.c(796): [client 2620:52:0:105f:5054:ff:fe22:d46f:44474] AH02415: [strict] Invalid host name '2620:52:0:105f:5054:ff:fe22:d46f', problem near: :52:0: [Wed May 10 04:41:31.930332 2017] [core:debug] [pid 15638] vhost.c(891): [client 2620:52:0:105f:5054:ff:fe22:d46f:44474] AH00550: Client sent malformed Host header: 2620:52:0:105f:5054:ff:fe22:d46f:8747 [Wed May 10 04:41:31.930338 2017] [core:debug] [pid 15638] protocol.c(1385): [client 2620:52:0:105f:5054:ff:fe22:d46f:44474] AH00569: client sent HTTP/1.1 request without hostname (see RFC2616 section 14.23): /
EAP6(6.4.10) log:
04:41:29,316 ERROR [org.jboss.modcluster] (ContainerBackgroundProcessor[StandardEngine[jboss.web]]) MODCLUSTER000042: Error null sending INFO command to 2620:52:0:105f:5054:ff:fe22:d46f/2620:52:0:105f:5054:ff:fe22:d46f:8747, configuration will be reset: null
Tomcat log:
May 10, 2017 9:05:25 AM org.jboss.modcluster.mcmp.impl.DefaultMCMPHandler sendRequest ERROR: MODCLUSTER000042: Error null sending INFO command to 2620:52:0:105f:5054:ff:fe22:d46f/2620:52:0:105f:5054:ff:fe22:d46f:8747, configuration will be reset: null
Issue was reproduced with EAP-7.1.0.DR17 too. Httpd-2.4.23-GA don't have this issue.
Issue has some connection to ipv6 as some ipv6 tests passed.
Failed on:
2620:52:0:105f:5054:ff:fe22:d46f, 2620:52:0:105f:5054:ff:fe0c:5717, 2620:52:0:105f:5054:ff:feb5:a642, ...
Passed on:
2620:52:0:1040:e8f4:2bff:feef:e504, 2620:52:0:1040:440c:3ff:fee1:df02, ...
Attachments
Issue Links
- is cloned by
-
-
- Resolved
-
