Host/Group replication between FOG Servers

[Wayne Workman]

@adukes40 The script from the main server looks good - that’s how it should look from the node.

Run them on the node. Yes, just copy paste.

[adukes40]

@Wayne-Workman this is what i get:

Sorry, I am new with Linux, so it is all a learning curve for me.

[Wayne Workman]

@adukes40 Try it with a simple password. It goes between the empty quotes.

[adukes40]

All I get is invalid syntax. No clue what I am doing wrong.

[Wayne Workman]

@adukes40 I sent you a message. Top right talk bubble.

[Wayne Workman]

I remoted in to help.

bind address 127.0.0.1 was set in my.cnf, I commented that and then we were able to connect to mysql from the remote node using the fogstorage username and password.

However, when visiting the storage node’s boot.php file with a registered MAC address appended like this:
http://10.106.2.149/fog/service/ipxe/boot.php?mac=b8:ac:6f:3d:6e:a4

It spits just this out:

set fog-ip
set fog-webroot
set boot-url http://${fog-ip}/${fog-webroot}

also, an apache2 error pops up in the node’s logs:
[Tue May 24 16:05:32.244071 2016] [:error] [pid 20780] [client 10.106.10.5:12079] PHP Fatal error: Call to a member function lastInsertId() on null in /var/www/html/fog/lib/db/pdodb.class.php on line 124

FOG version 7829
Both main and node is Ubuntu 14.04 LTS.

[Wayne Workman]

I just spoke briefly with the @Senior-Developers and they said that the storage node installation actually communicates with the main server’s DB to do certain things.

What this means is because of the bind address previously, those things the installer needed to do didn’t get done.

So, now that we’ve sorted out the connectivity issues, all you should need to do is re-run the installer on the storage node (no rebuild is required at all!), and then things should be working for you.

[Wayne Workman]

After re-running the installer on the storage node and rebooting the main server, this is what’s in the replication logs, and it just hangs and never replicates. @Tom-Elliott

[05-24-16 9:37:59 pm] | Image name: testing
[05-24-16 9:37:59 pm] * Found Image to transfer to 2 node(s)
[05-24-16 9:37:59 pm] | I am the only member
[05-24-16 9:37:59 pm] | Image Name: testing
[05-24-16 9:37:59 pm] * Not syncing Image between group(s)
[05-24-16 9:37:59 pm] | We are node name: CA - MasterNode
[05-24-16 9:37:59 pm] * We have node ID: #1
[05-24-16 9:37:59 pm] | We are group name: default
[05-24-16 9:37:59 pm] * We are group ID: #1
[05-24-16 9:37:59 pm] * Starting Image Replication.
[05-24-16 9:37:59 pm] * Starting service loop
[05-24-16 9:37:59 pm] * Checking for new items every 600 seconds
[05-24-16 9:37:59 pm] * Starting ImageReplicator Service
[05-24-16 9:37:59 pm] Interface Ready with IP Address: 167.21.42.13
[05-24-16 9:37:59 pm] Interface Ready with IP Address: 10.103.72.49

[Wayne Workman]

I’ve verified both the main and node’s FTP credentials.
I’ve toggled the master node for both, just to reset it.
I’ve made test images and test snapins but both hang at exactly where lftp should start.
I’ve also uninstalled and reinstalled lftp.

[Wayne Workman]

Found this interesting message in /var/log/syslog:
May 24 18:20:30 MSDCATS09 kernel: [ 2560.753652] init: vsftpd main process (4005) killed by TERM signal

[Wayne Workman]

snapins attempt to replicate, it seems… here’s the lftp log. But images are not.

root@MSDCATS09:/var/log# cat /var/log/xferlog
Tue May 24 17:44:36 2016 1 10.103.72.49 12 /opt/fog/snapins/test.bat b _ i r fog ftp 0 * c
Tue May 24 18:14:20 2016 1 10.103.72.49 12 /opt/fog/snapins/test.bat b _ i r fog ftp 0 * c

[Wayne Workman]

These log entries are also interesting:

cat /var/log/auth.log | grep vsftpd
May 24 17:36:01 MSDCATS09 vsftpd[21829]: pam_unix(vsftpd:auth): check pass; user unknown
May 24 17:36:01 MSDCATS09 vsftpd[21829]: pam_unix(vsftpd:auth): authentication failure; logname= uid=0 euid=0 tty=ftp ruser=apc rhost=10.103.67.62

[Wayne Workman]

@Tom-Elliott helped out, it was a bug in replication. It should be fixed in the current.