MAAS rack can't commission servers

Users
1

Hi,

I do have two servers racks in different Data Center . The first rack where the region+rack maas is installed are using the IP subnet 172.19.0.0/23 and is setup with ip 172.9.0.20, the second rack where the rack maas is installed has the subnet 172.16.92.0/24 and the IP is 172.16.92.249.

The problem is that I can not commission any machine . The error that I get is :
At machine :

cloud-init[1472]: request to http://172.16.92.249:5248/MAAS/metadata/2012-03-01/ failed. sleeping 1.: HTTP Error 504: Gateway Time-out
At maas rack :

[error] 116436#116436: *3700 upstream timed out (110: Unknown error) while reading response header from upstream, client: 172.16.92.51, server: , request: “POST /MAAS/metadata/2012-03-01/ HTTP/1.1”, upstream: “http://172.19.0.20:5240/MAAS/metadata/2012-03-01/”, host: "172.16.92.249:5248

But if I run the command

curl XXX.XXX.XXX.XXX:5248/MAAS/metadata/2012-03-01/

meta-data
user-data

Either at maas rack or at the machine .

I can run the network testing at machine and everything working fine .

P.S.: XXX.XXX.XXX.XXX is 172.16.92.254
Any help ?

2

Hi there,

MAAS version?

3

Hi MAAS version 3.3 (1:3.3.7-13246-g.2341ebf5d-0) installed from PPA .

4

I guess that your rack at 249 is forwarding the requests to the region 20 and for some reasons the region does not reply (maybe it was down, maybe a networking issue in your infra). This is expected to happen.

anything interesting in the .20 regiond logs at that time?

Can you still reproduce the issue on every commissioning operation?

5

Hi r00ta ,

What I can say about the network configuration :
I can ssh from 249 to 20 and also from 20 to 249 , that was my first test . All network is working because the syslog is working fine .
I could no see nothing at .20 log , any specific file would provide interesting information ?
At the .20 maas rack everything works fine , commission, deploy all good .

Now I figure out something that could be the problem . The connection between two Rack are limited to 1.0MB/s , I guess that can be the cause , right ? tested using a scp download .

We are working to fix that .

6

yes, this can be definitely a problem!

Keep me posted!

7

Problem solved , that was really related to network.
I don´t think that was only the poor performance but the traffic was not going correctly also .
Now we have commissioning working ,
Thanks r00ta for the help .

1 Like
8

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.