FogReplicator and Storage Nodes.

[Tom Elliott]

@george1421 True true true.

[sbenson]

@Tom-Elliott said in FogReplicator and Storage Nodes.:

True true true.

restarted on both
SRO

[17:05:16] root@SRO-FOG-01[0]:/var/log/fog$ systemctl restart FOGImageReplicator FOGSnapinReplicator; tail -n10 -f fogreplicator.log
[05-13-17 12:03:41 am]  * Found Image to transfer to 1 node
[05-13-17 12:03:41 am]  | Image Name: W7P-T460s
[05-13-17 12:03:42 am]  | W7P-T460s: No need to sync d1.fixed_size_partitions file to MHB-Slave
[05-13-17 12:03:42 am]  | W7P-T460s: No need to sync d1.mbr file to MHB-Slave
[05-13-17 12:03:43 am]  | W7P-T460s: No need to sync d1.minimum.partitions file to MHB-Slave
[05-13-17 12:03:43 am]  | W7P-T460s: No need to sync d1.original.fstypes file to MHB-Slave
[05-13-17 12:03:43 am]  | W7P-T460s: No need to sync d1.original.swapuuids file to MHB-Slave
[05-13-17 12:03:43 am]  | W7P-T460s: No need to sync d1.partitions file to MHB-Slave
[05-13-17 12:03:44 am]  | W7P-T460s: No need to sync d1p1.img file to MHB-Slave
[05-13-17 12:03:44 am]  | W7P-T460s: No need to sync d1p2.img file to MHB-Slave

MHB

[17:05:40] root@MHB-FOG-01[0]:/var/log/fog$ systemctl restart FOGImageReplicator FOGSnapinReplicator;tail -n10 -f fogreplicator.log
[05-03-17 4:01:10 am]  *  | This is not the master node
[05-04-17 4:11:10 am]  *  | This is not the master node
[05-05-17 4:21:10 am]  *  | This is not the master node
[05-06-17 4:31:11 am]  *  | This is not the master node
[05-07-17 4:41:11 am]  *  | This is not the master node
[05-08-17 4:51:11 am]  *  | This is not the master node
[05-09-17 5:01:11 am]  *  | This is not the master node
[05-10-17 5:11:11 am]  *  | This is not the master node
[05-11-17 5:21:11 am]  *  | This is not the master node
[05-12-17 5:31:11 am]  *  | This is not the master node

[george1421]

@sbenson The mhb looks a bit suspect, but is OK. For SRO I would expect this if MHB had already been synchronized on a previous run.

I do have to question the time differences. Are you running UTC on these systems, especially SRO?

[Tom Elliott]

@sbenson That is perfectly fine. The MHB Side, could probably do with a (ctrl+c) then rerun the tail command.

Notice how the restart didn’t have the “banner”?

[sbenson]

ok…good news…it’s replicating an image…that I dont want it to(it was set to replicate). So i know replication is working now. Now to test if I deploy a machine from MHB to see if it hits SRO at all… I will keep you updated. Need to reach out to someone in MHB

[sbenson]

@Tom-Elliott I didn’t realize the restarting of the service deleted the file and recreated it. I thought it would just append to it

[Tom Elliott]

@sbenson It’s kind of intentional to ensure a “clean state”.

[Wayne Workman]

@sbenson said in FogReplicator and Storage Nodes.:

@Tom-Elliott I didn’t realize the restarting of the service deleted the file and recreated it. I thought it would just append to it

Yeah, historical replicator logs aren’t important. What’s important are the current logs it’s writing.

[sbenson]

Thank you everyone, this has been resolved. I will say the replication/node setup was the most confusing thing about fog. Everything else is straight forward. Maybe it was just me but a “Storage Nodes for Noobs” might be a good sticky post that explains how to setup a simple 2 server storage node. Other than that everything has been awesome. Thanks for all your hard work.

[Wayne Workman]

@sbenson said in FogReplicator and Storage Nodes.:

Maybe it was just me but a “Storage Nodes for Noobs” might be a good sticky post that explains how to setup a simple 2 server storage node.

I agree. We already have documentation on replication and the location plugin but we need an instructional video on setting up a simple two-node system. I can get a video made pretty easily on this topic, and I’ve been looking for simple topics to make videos about.

[george1421]

@Wayne-Workman Just be aware the OP has setup a FOG Master -> FOG Slave setup not the traditional and fully supported FOG Master -> [FOG] Storage node configuration.

My hope is as the API matures that we can have the FOG Master -> FOG Slave setup a supported configuration.

[Wayne Workman]

@george1421 I didn’t know this was a multi-master setup, if that’s what you meant. I was referring to a video on just configuring a second node into an existing fog setup. With the new access control plugin and site plugin, multi-master setups are just that less a good idea.

As it stands now with the access control plugin, location plugin, and site plugin - the only benifit one stands to gain in a multi-master setup is some level of fault tolerance against the upstream box failing - and fault tolerance against the link failing - and for those two things you loose a comprehensive database/control center in a central location - a major drawback.