Web Hosting Talk : Web Hosting Main Forums : VPS Hosting : Installed Openz slave can't ssh into a vps

[captain_squash]

I've been trying to figure this out for a solid 5 hours now and I'm at my wits end so I thought I would post here and hopefully find some answers.

I have purchased a new dedicated server and installed an OpenVZ Slave for SolusVM.

Everything appears to install fine, I've done this twenty times with no issues. The only difference with the current server is I'm trying a new provider.

When I add a vps everything also appears to go fine. In SolusVM it shows the vps running.

However I cannot ssh into the vps. I can ping the ip but if I try:

telnet xxx.xxx.xxx 22 I just get "Could not connect etc.."

I can ping the vps from the host node and if I login to the vps via the SolusVM console I can ping the host node. However on the vps I cannot resolve any domains.

I've checked my settings in: http://wiki.openvz.org/Quick_installation#sysctl and everything seems fine.

If anyone could offer any assistance it would really be appreciated.

If more details are needed, please let me know.

[CN-Jeremy]

You might try downgrading the version of vzctl.. I know in the past it caused issues if a newer version of vzctl was used.

http://wiki.solusvm.com/index.php?title=OpenVZ_Bugs

Not sure if this still applies since the recent updates but your issue sounds similar to one I ran into a while back and basically it was not setting the gateway correctly. If I manually set the gateway in the vps via console it would work until I rebooted the vps.

Hope this helps.

[kbeezie]

Can you SSH into the OVz Slave node, then use vzctl enter ### where ### is the number of the VPS container (i.e.: do vzlist to see what it's number is) and then check to see if SSHd is even running on the VPS you are running. Can probably even vzctl restart ### to see if that does the trick either.

[captain_squash]

Thanks for the replies, the vzctl version is the name version I'm using on other nodes so I trust it works.

I entered the vps from the node and checked sshd and it appeared to be running.

However when I restarted the node I got the following error:

[root@108 private]# vzctl restart 1861
Restarting container
Stopping container ...
Container was stopped
Container is unmounted
Starting container ...
Container is mounted
Adding IP address(es): xxx.xxx.xxx.xxx xxx.xxx.xxx.xxx
arpsend: xxx.xxx.xxx.xxx is detected on another computer : 00:0f:f8:41:fc:00
vps-net_add WARNING: arpsend -c 1 -w 1 -D -e xxx.xxx.xxx.xxx eth0 FAILED
arpsend: xxx.xxx.xxx.xxx is detected on another computer : 00:0f:f8:41:fc:00
vps-net_add WARNING: arpsend -c 1 -w 1 -D -e xxx.xxx.xxx.xxx eth0 FAILED
Setting CPU limit: 300
Setting CPU units: 1000
Setting CPUs: 3
Container start in progress...

Getting ready to throw in the towel at this point.

[ozfingwe]

it looks as if another server (or VPS) has the VPS' IP in question assigned, one that allows icmp but not ssh. Are you on your own subnet or using a range of a subnet only? Checked ifconfig -a and did a traceroute to the IP that has issues?

cheers
oz.

[captain_squash]

Originally Posted by ozfingwe

it looks as if another server (or VPS) has the VPS' IP in question assigned, one that allows icmp but not ssh. Are you on your own subnet or using a range of a subnet only? Checked ifconfig -a and did a traceroute to the IP that has issues?

cheers
oz.

I asked the DC and got the following answer:

"You are on a dedicated vlan, the ip isn't in use in our datacenter. Sounds like you already have the ip bound to your server."

(I'm honestly not sure about the subnet question)

[kbeezie]

It's possible that the host node has the full IP range assigned and bound, as opposed to only having the first IP bound (thus leaving the rest usable by any new VPS)

When you do a ifconfig from the slave node, do you see a full range of IP addresses or just your first one? The error seems to imply that you have them all pre-setup on the slave node as it mentions eth0 as the possible interface already having it bound.

[captain_squash]

Here's the output of my ifconfig, I hope this helps:

[root@108 ~]# ifconfig
eth0 Link encap:Ethernet HWaddr 54:04:A6:3C1:1D
inet addr:108.171.212.210 Bcast:108.171.212.223 Mask:255.255.255.224
inet6 addr: fe80::5604:a6ff:fe3c:d11d/64 Scope:Link
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:42563 errors:0 dropped:0 overruns:0 frame:0
TX packets:15779 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:3430136 (3.2 MiB) TX bytes:2636661 (2.5 MiB)
Interrupt:130 Memory:fb400000-fb420000

lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
inet6 addr: ::1/128 Scope:Host
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:19 errors:0 dropped:0 overruns:0 frame:0
TX packets:19 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:2418 (2.3 KiB) TX bytes:2418 (2.3 KiB)

venet0 Link encap:UNSPEC HWaddr 00-00-00-00-00-00-00-00-00-00-00-00-00-00-00 -00
inet6 addr: fe80::1/128 Scope:Link
UP BROADCAST POINTOPOINT RUNNING NOARP MTU:1500 Metric:1
RX packets:4058 errors:0 dropped:0 overruns:0 frame:0
TX packets:23 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:215314 (210.2 KiB) TX bytes:1863 (1.8 KiB)

Thanks again for the replies.

[kbeezie]

And the IP you're attempting to add/use with the VPS is one that belongs to that physical machine (and is not the primary IP listed there) ?

[captain_squash]

Originally Posted by kbeezie

And the IP you're attempting to add/use with the VPS is one that belongs to that physical machine (and is not the primary IP listed there) ?

Yeah I didn't use the primary ip listed there for the vps.

Funny thing is it doesn't matter what ip I use from the block I was given, they all give me the same issue.

[kbeezie]

Originally Posted by captain_squash

Yeah I didn't use the primary ip listed there for the vps.

Funny thing is it doesn't matter what ip I use from the block I was given, they all give me the same issue.

It does matter if the IP you're trying to use was not assigned to that physical slave server. i.e.: You could have 2 physical servers (I don't know if you're just using one or not, but usually you have 2 or more if you're installing an Ovz Slave node), and if you tried to migrate a VPS from Node #1 to the Slave node and did not update it's IP, it may be trying to use an IP that isn't assigned to that physical node.

[captain_squash]

Originally Posted by kbeezie

It does matter if the IP you're trying to use was not assigned to that physical slave server. i.e.: You could have 2 physical servers (I don't know if you're just using one or not, but usually you have 2 or more if you're installing an Ovz Slave node), and if you tried to migrate a VPS from Node #1 to the Slave node and did not update it's IP, it may be trying to use an IP that isn't assigned to that physical node.

It's a new node I bought a few days ago. They assigned me a block of ips for the node and I figured they would work just fine. Basically every time I create a test vps on the node it is the first time I am using the particular ip from the block.

[kbeezie]

Originally Posted by captain_squash

It's a new node I bought a few days ago. They assigned me a block of ips for the node and I figured they would work just fine. Basically every time I create a test vps on the node it is the first time I am using the particular ip from the block.

So:

New Node -> Specific block of IPs they gave you for that exact node -> Trying on VPS

Now how are you going about creating the VPS? Did you make sure to specify the new node in the drop down when adding the VPS and you made sure to add the IP block in solusvm assigned to the new node as well?

[captain_squash]

Originally Posted by kbeezie

So:

New Node -> Specific block of IPs they gave you for that exact node -> Trying on VPS

Now how are you going about creating the VPS? Did you make sure to specify the new node in the drop down when adding the VPS and you made sure to add the IP block in solusvm assigned to the new node as well?

Yes that is correct

I've actually setup 20 other slaves before and this is the first time I've run into this issue. /stumped

[kbeezie]

Originally Posted by captain_squash

Yes that is correct

I've actually setup 20 other slaves before and this is the first time I've run into this issue. /stumped

Hrm... weird. I'm assuming the new node is pretty empty other than the VPS you're trying to create, maybe try to bind something (say a simple web process or some quick python app) on the IP and see if you still get the problem, least it would verify it's not just a OpenVz thing.