Can't get FOG to replicate to storage node
-
Sorry for the delay in posting. Since my last post, I did something to my database and locked myself completely out. I purged everything and reloaded. My main fog server is running great (actually better, I had another unrelated issue that is now gone) and I just installed the storage node and followed step by step instructions on the web and got nothing. I retried the instructions to try to connect to the storage node through the main server
[CODE]first off - run the below command on your main server:
mysql -u fog fog -pPASSWORD_GIVEN_AFTER_INSTALLING_FOG_ON_THE_NODE -hIP_OF_NODE[/CODE]and once again got
[CODE]ERROR 2003 (HY000): Can’t connect to MySQL server on ‘10.24.1.2’ (111)
[/CODE]I commented out the bind-address line on my node and restarted with no change.
-
I have found that the 13.10 Ubuntu revision has issues when installing fog.
I had to manually restart the Mysql and apache2 services after a complete install (the install reported no errors, however when trying to access the Web GUI or the database I received error messages).
After manually enabling the service I was able to use the server.
I Would still spend some time trying to troubleshoot the issue incase you can get it working, or incase the issue actually lies else where, Lee’s troubleshooting steps are recommended.
In the even that you are unsuccessful, if you’re not opposed, a complete wipe and re-install would not be a bad idea. After completing the ./installfog.sh type
[code]
sudo service mysql restart
sudo service apache2 restart
[/code] -
Jaymes, I have completely reinstalled both my main FOG and the storage node and started from scratch. This did resolve a different issue with searching for hosts but did not resolve the issue of communication between my main and node.
-
The communication issue, I think, is because of two possible (maybe even both) scenarios.
First things first, I’m assuming you’ve removed the bind-address = 127.0.0.1 from the my.cnf files on both storage-node and server as Lee stated above?
Second, did you create the fogstorage user on both the server and the node with the same username and password?
Are both system without firewall?
Next,
When you created the FOGUser, did you just create them as ‘fogstorage’@‘localhost’?
For now I would attempt using the root user and password for both as you can troubleshoot the fogstorage user later on. We want to see both the node connect to the server AND the server connect to the node. This way you know communication happens between both.
-
I removed the bind address on both server and node, and both systems do not have a firewall. I checked mysql on the server and made sure the password was set. On node, mysql-server is not installed so there is no user to create or check.
I have set the node to try using root user, which I can log into the server from node with, and restarted FOGImageReplicator and Im still getting.
[CODE][02-27-14 7:32:54 pm] * Starting FOG Image Replicator Service
[02-27-14 7:32:59 pm] * [02-27-14 7:32:59 pm] Checking for new tasks every 600 seconds.
[02-27-14 7:32:59 pm] * [02-27-14 7:32:59 pm] Starting service loop.
[02-27-14 7:32:59 pm] [02-27-14 7:32:59 pm] Failed to connect to database server, will try again in next iteration.[/CODE] -
Looking back over the posts still trying to understand what is needed. I looked at Lee’s post, troubleshooting connections between the two machines and realized something, there is no Mysql database on the node mysql-server does not install when installing a node.
-
After an experience I had yesterday, I don’t think a storage node is what I am wanting to use. I was pushing out an image to a machine and noticed that it was connecting to my storage node to try to pull the image (it errored out because the image still has not replicated.)
The problem I have with this is that my intent was to have a master node at my main school and a storage node at various campuses but what I found yesterday was that if I am imaging a machine at campus A it may pull an image from Campus B which would take up precious bandwidth.
Someone correct me if I am wrong but what I believe the storage nodes are most useful for is multi-casting and in my particular set-up it would be better to have individual master nodes at each campus and copy any new images manually to each campus.
-
Thats exactly why i devloped the location patch and Tom has now recoded it as a plugin. You set the location as Campus A for the host and it will only pull the image from that node that uve assigned to location campus A and if u dnt slecify a node u can set it to a storage group also.
-
Good to know. I did some searching around and it looks like the patch is available for .31 and will be in .33 but currently not in .32. Is that right?
-
Yes i never released the re-coded version for 0.32 and it already is available in 0.33b
-
When will .33 be available for upgrade?
