| loopback ipv4 tcp not working [message #52103] |
Thu, 25 June 2015 20:08  |
tramsdell
Messages: 1
Registered: June 2015
|
Junior Member |
|
|
Hello.
There are a few old postings on this topic but the suggested fixes are don't have any effect. Hopefully some of this is new info.
Essentially, the container behaves as though there is firewall blocking TCP over IPv4 in one particular container. The following commands do work as expected:
ssh localhost // this connects successfully through the IPv6 loopback address
ping 127.0.0.1 // ping replies normally on the IPv4 address and is observable with tcpdump
This does not work:
ssh 127.0.0.1
Tcpdump shows the IPv4 SYN packets but there is no inbound traffic at all. This is not peculiar to ssh, anything trying to communicate TCP over the IPv4 loopback does not work. Sniffing ssh to the IPv6 address shows the expected traffic. I've not tried UDP.
Based on other posts, the ip_forward was set to 1 with no effect. Another post had 127.0.0.1 assigned to multiple interfaces - this is not the case here. iptables is not running. Other containers on the same host do not have this problem and appear to be configured identically to the broken one.
For what it's worth, the process we run in the container crashed and management software restarted the container. After the restart, we have this loopback problem.
Are there any logging or diagnostics that can be enabled in this area? There are no obvious messages in /var/log/messages. Any other suggestions?
Thanks.
Tom
|
|
|
|
| Re: loopback ipv4 tcp not working [message #52104 is a reply to message #52103] |
Sun, 28 June 2015 12:41  |
Paparaciz
Messages: 302
Registered: August 2009
|
Senior Member |
|
|
on what ip addresses services are listening to? is container have ipv4 address or only ipv6?
please also report HN os, kernel and openvz tools versions.
please report container os and version.
please show netstat -an, ifconfig -a output from the container.
|
|
|
|
|
|
| Re: loopback ipv4 tcp not working [message #52192 is a reply to message #52104] |
Mon, 07 September 2015 04:04 |
happyfunball
Messages: 18
Registered: November 2007
Location: Toronto
|
Junior Member |
|
|
We see this issue from time to time as well, seems to be related to the number of containers on a node. No failcnt is triggered.
Loopback connections (127.0.0.1) randomly time out and we don't see a reason in the logs. We can confirm this problem in kernel 042stab108.8
[Updated on: Mon, 07 September 2015 04:05] Report message to a moderator |
|
|
|
OpenVZ Forum
Members
Search
Help
Register
Login
Home
