[Users] Some VE on one HW node cannot communicate each other (getting Time to live exceeded)
Hello, I've server with two network cards in following setup: auto eth0 iface eth0 inet static address 195.113.233.254 netmask 255.255.255.0 network 195.113.233.0 broadcast 195.113.233.255 gateway 195.113.233.1 auto eth0:0 iface eth0:0 inet static address 192.168.1.1 netmask 255.255.255.0 network 192.168.1.255 broadcast 192.168.1.255 auto eth1 iface eth1 inet static address 195.113.187.37 netmask 255.255.255.192 network 195.113.187.0 broadcast 195.113.187.63 post-up ip rule add from 195.113.187.33 table 6 post-up ip rule add from 195.113.187.35 table 6 post-up ip route add default dev eth1 via 195.113.187.1 table 6 post-down ip rule delete from 195.113.187.33 table 6 post-down ip rule delete from 195.113.187.35 table 6 There are three networks on: 1) eth0 195.113.233.0/24 primary public network for OpenVZ systems 2) eth0:0 192.168.1.0/24 private network for system which should not waste publicaly routable IP but still need to be online 3) eth1 195.113.187.0/26 is secondary public network - I was using this range for physical systems which I'm now moving to VE Everything is working fine, except comunication between systems using eth1 and eth0. Routing tables on HW node: chlivek:~# ip rule 0: from all lookup 255 32764: from 195.113.187.35 lookup 6 32765: from 195.113.187.33 lookup 6 32766: from all lookup main 32767: from all lookup default chlivek:~# ip route 195.113.187.33 dev venet0 scope link src 192.168.1.1 192.168.1.2 dev venet0 scope link src 192.168.1.1 195.113.233.253 dev venet0 scope link src 192.168.1.1 195.113.233.252 dev venet0 scope link src 192.168.1.1 195.113.187.0/26 dev eth1 proto kernel scope link src 195.113.187.37 195.113.233.0/24 dev eth0 proto kernel scope link src 195.113.233.254 192.168.1.0/24 dev eth0 proto kernel scope link src 192.168.1.1 default via 195.113.233.1 dev eth0 chlivek:~# route -n Kernel IP routing table Destination Gateway Genmask Flags Metric RefUse Iface 195.113.187.33 0.0.0.0 255.255.255.255 UH0 00 venet0 192.168.1.2 0.0.0.0 255.255.255.255 UH0 00 venet0 195.113.233.253 0.0.0.0 255.255.255.255 UH0 00 venet0 195.113.233.252 0.0.0.0 255.255.255.255 UH0 00 venet0 195.113.187.0 0.0.0.0 255.255.255.192 U 0 00 eth1 195.113.233.0 0.0.0.0 255.255.255.0 U 0 00 eth0 192.168.1.0 0.0.0.0 255.255.255.0 U 0 00 eth0 0.0.0.0 195.113.233.1 0.0.0.0 UG0 00 eth0 Pinging from system 195.113.187.33 outside: ermon:~# ping -c 1 195.113.233.254 (HW node) PING 195.113.233.254 (195.113.233.254) 56(84) bytes of data. 64 bytes from 195.113.233.254: icmp_seq=1 ttl=64 time=0.035 ms ermon:~# ping -c 1 195.113.233.1 (Gateway) PING 195.113.233.1 (195.113.233.1) 56(84) bytes of data. 64 bytes from 195.113.233.1: icmp_seq=1 ttl=254 time=0.472 ms ermon:~# ping -c 1 195.113.233.253 (VE using eth0) PING 195.113.233.253 (195.113.233.253) 56(84) bytes of data. From 192.168.1.1 icmp_seq=1 Time to live exceeded I'm bit curious where that 192.168.1.1 get into path. In FW rules I'm using: iptables -t nat -A POSTROUTING -s 192.168.1.0/24 -o eth0 -j MASQUERADE echo 1 /proc/sys/net/ipv4/ip_forward for translating 192.168.1.0/24 onto public 195.113.233.254. I used http://wiki.openvz.org/Source_based_routing as base for setup, but I'm really clueless how to fix it. Please can someone review this setup and suggest me better way how to setup this? Best regards -- --- Jan Tomasek aka Semik http://www.tomasek.cz/ signature.asc Description: OpenPGP digital signature ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Installing BIND on VE's
On 12/21/06 9:19 PM, Al Sparks [EMAIL PROTECTED] wrote: I'm trying to set up a test environment of name servers. I downloaded the BIND tarball to one of the Virtual Environment, and tried to compile it. That didn't work, because there's no gcc installed. So I tried compiling it at the management level which did work, and transferred the compiled version over to the VE. When I did a make install make did run, but ended with errors. How do I create a VE that will compile and install something like BIND? I'm sure there are lots of libraries that may be needed. I'm new to OpenVZ. Am I missing something obvious? I'm running CentOS 4.4 in all VE's. === Al ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users Al, the first thing you want to do is install yum in your vps from the host node. You would do that from the Host Node with something like this: vzyum vpsid install yum vpsid being the id of the vps you want to install yum in. Then I would ssh into the vps and using yum install BIND or at least the libraries you will need to compile BIND. You would do that with something like this: yum install bind HTH's Romeo ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Installing BIND on VE's
Jan Tomasek wrote: Hello Al, Al Sparks wrote: I'm trying to set up a test environment of name servers. I downloaded the BIND tarball to one of the Virtual Environment, and tried to compile it. That didn't work, because there's no gcc installed. why you do not install gcc into VE? I'm handling VE almost as ordinary system running on physical HW. Template which come with OpenVZ (at least for Debian which I'm using) is very basic instalation of Linux, I have to install there numerous packages to be able use it for my work. That's why it is called minimal -- it's a boilerplate, so you should use apt-get/aptitude/any-other-tool to install whatever you need on top of that. By the way you can tar the result and use it as a template to create other VEs. ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] DNS Server on OpenVZ.
On Friday 22 December 2006 18:29, Paul Stuffins wrote: the node. I have no problems installing OpenVZ and creating VE's but I am under the impression that a VE can not see any other VE's on the same node, is this correct? I ask because one of the VE's will be The VEs send all their traffic to the Hostnode. So by default the VEs can communicate with the HN only. For any other kind of communication you have to enable IP-Forwarding on HN - usually one does this together with Proxyarp. By doing this the VEs can communicate quite well because the HN does all the routing. So there should be no problem with your plans. For further information fire up a search-engine and ask for Pseudobridge with proxyarp. There is a howto somewhere that covers this. Cheers, Konstantin ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users
Re: [Users] Installing BIND on VE's
Al, the first thing you want to do is install yum in your vps from the host node. You would do that from the Host Node with something like this: vzyum vpsid install yum vpsid being the id of the vps you want to install yum in. Then I would ssh into the vps and using yum install BIND or at least the libraries you will need to compile BIND. You would do that with something like this: yum install bind Thanks for all the responses. I didn't install yum on the VE. Instead, I used vzyum from the host mode to install gcc, flex, byacc, which is what Bind-8 demands to do an install. Thanks for the help. === Al ___ Users mailing list Users@openvz.org https://openvz.org/mailman/listinfo/users