Save changes to aerospike.conf

Hi Guys, I’m Sairam I’m trying to edit the aerospike.conf file using this command

[vagrant@vm1hn ~]$ vim /etc/aerospike/aerospike.conf

I have entered i to insert and this warning got displayed – INSERT – W10: Warning: Changing a readonly file

After, I made configuration changes but, I was unable to save the configuration and it says

E45: ‘readonly’ option is set (add ! to override) 29,10-24 Top

Could any one please guide me. I appreciate your help. Thanks for your time

You are trying to edit the config with a user that doesn’t have access to the file. Typically you would use sudo.

sudo vim /etc/aerospike/aerospike.conf

Read more about it here: Sudo - ArchWiki

Thanks Kporter it helped to save changes in aerospike.config file. Now, I have new issue when you have time could you please help me.

when I’m trying to add nodes suddenly I have a warning called connection lost and when I try to reload it to add nodes this page got displayed.

Here is my vagrantfile

AMC could just be down. Try recycling AMC or check cluster health using asadm

Adding a fresh node cannot be done through AMC. The node has to be configured correctly, and then brought up and it will join the cluster and finally will show on AMC.

Thanks, Meher, and How should I know that vagrantfile is configured correctly. You said the node has to configure correctly it means that I should edit the aerospike.config file in all the three VM’s, right?

You are correct.

Thanks, Meher. I have three VM’s running in vagrant and I have given the 127.0.0.1 to add the cluster in AMC and I see three nodes.

I haven’t made any changes to the aerospike.config file and what I understood was the cluster I have added has three nodes but they don’t communicate with each other.

Let’s say I need a replica of data 3, and when one node fails the other node, will use partition map and make sure that replica of data 3 is maintained with 2 nodes.

But, when I don’t configure the aerospike.config file the above-mentioned story will not happen.

Hi Meher, This is my understanding please correct me if I’m wrong. Thanks for your time and I appreciate your help.

Not sure I understand what you wrote. From the screenshot, you seem to have a 3 node cluster, and if you haven’t changed any configuration, it means that it is working through multicast and the nodes would be communicating with each other. If you have replication factor 3 on a 3 node cluster and lose one node, you can then only have replication factor 2 on the remaining 2 nodes.

Don’t I need to make any changes to the config file to achieve 3 node cluster multicast mode

these were the screen shots of 3 vm’s config files

If you keep the default config and multicast is enabled on your network, it will just work as it seems to have. I would encourage you to read the documentation details on how to setup / configure / manage Aerospike. For this particular topic, the details are here: http://www.aerospike.com/docs/operations/configure/network/heartbeat.

Hi Meher,

I’m running all the three VM’s, asd, amc where also running fine but all the three nodes where not displayed in web browser amc.

could you please help me find the issue here.

I would recommend checking the logs of each of your nodes to figure out why a 3 node cluster was not formed. You obviously managed to have a 3 node cluster formed previously, or so it seems, from your previous screen shot.

Yes, you are right previously, I have formed 3 three-node cluster but when I tried again I’m facing this issue. How do I check the log files? I have entered the VM1

Which command should I use here?

The log file is by default under /var/log/aerospike/aerospike.log

VM1

VM2 VM3

Only VM1 is showing in AMC, VM2, VM3 is not there.

I couldn’t understand the problem here ? Please help me.

The nodes don’t seem to properly see each other. My guess is that the traffic over fabric is not properly working… make sure your nodes can see each other over the fabric port 3001, through the interface specified (if any) under the fabric stanza. But it is hard to debug with limited logs.