I am setting up a grid-enabled cluster. I plan to assign 2 IP to my head node: one for local connection (LAN for distributing jobs to compute nodes) and one for public (internet for user access). So, my /etc/hosts file looks something like this:
111.111.111.111 myserver.whatever.com myserver #for public IP
11.11.11.11 myserver.whatever.com myserver #for local LAN
22.22.22.22 computenode01
33.33.33.33 computenode03
My concern here is will the hostname of myserver get messed up since it is mapped to two IPs?
I fear the system will always choose the first entry (111.111.111.111) if you want to resolve "myserver" address.
It will simply ignore the second entry, as I guess. Choose different hostnames for each entry, e.g. myserver.local and myserver.remote.
Related
I have a bare metal server, and I want to install multiple services on this server.
My inventory looks like that
[Mygroup]
Server port_service=9990 service_name="service1"
Server port_service=9991 service_name="service2"
When I launch my ansible job,only service 2 is installed because I have the same server in each line of my group. There is way to force ansible to take all lines of a group?
I don't want to create a group for each service
Q: "There is a way to force Ansible to take all lines of a group?"
A: No. There is not. In a group, the hosts shall be unique. If there are multiple hosts with the same name the last one will be taken.
Put the variables into one line e.g.
[Mygroup]
Server port_services="[9990, 9991]" service_names="['service1', 'service2']"
(and change the code).
See How to build your inventory. There is a lot of other options e.g.
[Mygroup]
Server
[Mygroup:vars]
port_services="[9990, 9991]"
service_names="['service1', 'service2']"
I hope I got u right but this should do the trick.
https://docs.ansible.com/ansible/latest/user_guide/intro_patterns.html
Greets
Harry
Another solution is to use an alias.
This solution works fine for me
[Mygroup]
service_1 ansible_host=Server port_service=9990 service_name="service1"
service_2 ansible_host=Server port_service=9991 service_name="service2"
I have a tower Template which when launched prompts users to provide hostname and it's IP separated by space per line. There will always be minimum two entries.
hostnameX ip_address
hostnameY ip_address
I am using the IP address and hostname to create DNS records in DNS servers in another task.
But for other task I just need the hostname and discard the IP address, and loop through the hostnames. But when I try
with_list: "{{list_serverinfo.split(' ')[0] }}"
The task is only run on hostnameX and never executes on hostnameY! I want the task to execute against all hostnames.
Thanks in advance for the time and all the help!
I can't seem to extend my pool for my additional users. I don't have any problem with my connection when the ip address handed is 192.168.10.xxx but when it reaches to 192.168.11.xxx to 192.168.16.xxx, it can no longer use the internet. What am I missing with my setup?
Updated:
If you want to add more IPs in your DHCP server, just increase the range of your current network. You currenlty have 253 clients (192.168.10.0/24)
Don't add more /24 networks, it's useless, just use for example 192.168.8.0/21 (range .8.1 to .15.254) to get 2046 IPs. See http://www.subnet-calculator.com/subnet.php to test various network ranges.
So I suggest this:
remove parasites /24 networks and NAT rules (192.168.11.0/24 to .16/24)
increase the range of your current network: change IP address from 192.168.10.1/24 to 192.168.10.1/21, subnet 255.255.248.0, network 192.168.8.0
change NAT/masquerade rule: src-address=192.168.8.0/21
change dhcp network range 192.168.8.0/21
change dhcp pool size with two segments: 192.168.8.1-192.168.9.254 and 192.168.10.100-192.168.15.254
and, normally, it should work
Here is one deployment scenario wherein NAT exists between DC boundaries, following are the requirements:
Cassandra Version: 2.1.13
There are 2 DCs, two Cassandra nodes (dc1:node1 & dc2:node3) across DCs should communicate across NAT boundaries using public IP.
One of the DC which is behind NAT has 2 Cassandra nodes (dc1:node1 and dc1:node2) and both them should communicate within NAT using private IP.
All these 3 nodes (dc1:node1, dc1:node2 & dc2:node3) should form a ring and communicate with each other.
Looked into seeds, listen_address, broadcast_address & broadcast_rpc_address.
https://docs.datastax.com/en/cassandra/2.1/cassandra/configuration/configCassandra_yaml_r.html
If there are public IPs used in broadcast_address and seeds then across DC & NAT communication works, however the nodes which are within NAT not able to detect each other.
If there are private IPs used in broadcast_address and seeds then within DC & NAT communication works, however the nodes across DC & NAT not able to detect each other.
Looked into Ec2MultiRegionSnitch but that will not work for premise deployments: https://docs.datastax.com/en/cassandra/2.1/cassandra/architecture/architectureSnitchEC2MultiRegion_c.html#architectureSnitchEC2MultiRegion_c__other-settings
What configuration settings will be required to achieve above 3 requirements?
Use gossiping property file snitch, set public ip as broadcast_address and private ip as listen_address. They will use the listen_address in same DC and broadcast address if in other DC.
Settings made with seeds=public address, listen_address=private and broadcast_address=public.
With these settings made on all 3 nodes:
dc1:node1 & dc2:node3 works but
dc1:node1 and dc1:node2 doesn't
Is it because of seeds have public address within DC behind NAT might nor work?
One of the observation to check listen on dc1:node1, private IP is listed:
node1# netstat -anp | grep -E "(7001)"
tcp 0 0 dc1:node1_privateIP:7001 0.0.0.0:* LISTEN 9999/java
Can dc1:node2 establish connection with dc1:node1_publicIP?
Is this https://issues.apache.org/jira/browse/CASSANDRA-9748 related here or will be only applicable in case of multiple NICS and not NAT environment?
I have a server with multiple public IP addresses.
I want to send campaign emails on this server.
Sometimes i would like to send mail from a particular IP (it is a filter on the sender email address that gives which IP to use).
The only thing i find is to install multiple postfix instances (one per output IP). Is there a best way to do this ?
I have a second question: Postfix gives a unique queue id to each message. If i have several instances of postfix, do you think thoses uniques id can be the same in 2 postfix instances ?
Thanks
sender_dependent_default_transport_maps is your friend. First, add this to main.cf:
sender_dependent_default_transport_maps = hash:/etc/postfix/sender-transport
Next, create the file /etc/postfix/sender-transport with
#my-sender-domain.com smtp-192-168-0-1:
Any message received with sender #my-sender-domain.com will use the service smtp-192-168-0-1 (can be any name) for sending. Don't forget to postmap /etc/postfix/sender-transport the file.
And then, add the service to master.cf
smtp-192-168-0-1 unix - - n - - smtp
-o smtp_bind_address=192.168.0.1
Again, the service name can be anything, but it must match the one on the hash file. This smtp service will send the message from the IP 192.168.0.1. Change as needed.
Add as many services and lines in the hash file as you want. Don't forget to service postfix restart after that.
There are many other options you can add to the smtp service, like -o smtp_helo_name=my.public.hostname.com, etc.
I just finished to set up a postfix like this :-)