Image Replication Across Storage Groups



  • Hi All,
    I have got 4 office locations and, at the moment, 2 FOG instances. One master at location A and in storage group 1 and one storage node in location B win storage group 2.

    what i want to end up with is a storage group for each location with a storage node in that location (Physically and in FOG config) but the single master in location A. This would then allow image management for all remote storage node locations from my base office location.

    Is that possible?

    Thanks,
    Matt.



  • @coxm said in Image Replication Across Storage Groups:

    should i image a pc in that remote site which is made part of that remote site storage group it will pull the image from that node, not the master Fog server on my local site.

    Not necessarily. Without the location plugin, fog just picks whatever server that has empty slots.

    You should look at the location plugin (link posted earlier), I think that’s what you need. The location plugin allows you to assign storage groups and storage nodes a physical location, and allows you to set physical location on hosts. If host A is assigned location A, and a storage node is assigned location A, then all imaging on host A would only use node A.



  • @wayne-workman Thanks for that.
    I am trying to confirm that now I have an image replicated to a storage node on my remote site, should i image a pc in that remote site which is made part of that remote site storage group it will pull the image from that node, not the master Fog server on my local site.

    Hopefully that all makes sense!



  • @coxm said in Image Replication Across Storage Groups:

    Just to confirm, any machine I put into the same storage group as the storage node will pull the data from that box rather than the Full Fog master server, is that correct?
    Thanks.

    I’m not sure I understand your question, so I’ll explain the order of operations.

    Replication first does group-to-group which is from a master group’s master node —> non-master group’s master node.
    Next, replication does inner-group replication from a group’s master node to all non-master nodes in the same group.

    This is outlined in the wiki article on replication.



  • @Wayne-Workman Dont worry, was a Linux permission issue on the 2nd data disk. All sorted and replicating as we speak!
    Thanks.

    Just to confirm, any machine I put into the same storage group as the storage node will pull the data from that box rather than the Full Fog master server, is that correct?
    Thanks.
    Matt.



  • @wayne-workman
    Thanks, looks like thats all sorted. The problme seems to now be the file path specified on the Storage node
    Also, can you confirm what the “deleting file…” is about, that image is in use on live systems so dont need that disappearing!

    mirror: Access failed: 550 Failed to change directory. (/media/ubuntuadmin/Data/Images/Win10-Desktop)
    [12-14-17 4:01:19 pm] | There are no other members to sync to.
    [12-14-17 4:01:19 pm] | Image Name: Win10-Desktop
    [12-14-17 4:01:19 pm] * Not syncing Image between nodes
    [12-14-17 4:01:19 pm] * Attempting to perform Group -> Nodes image replication.
    [12-14-17 4:01:19 pm] * Started sync for Image Win10-Desktop
    			lftp -e 'set xfer:log 1; set xfer:log-file "/opt/fog/log/fogreplicator.Win10-Desktop.transfer.ATL-FOG.log";set ftp:list-options -a;set net:max-retries 10;set net:timeout 30; mirror -c -r -R --ignore-time -vvv --exclude ".srvprivate" "/images/Win10-Desktop" "/media/ubuntuadmin/Data/Images/Win10-Desktop"; exit' -u fog,[Protected] 10.0.0.23
    [12-14-17 4:01:19 pm] | CMD:
    [12-14-17 4:01:19 pm] * Starting Sync Actions
    [12-14-17 4:01:18 pm] * Deleting remote file: 
    [12-14-17 4:01:18 pm] | Files do not match.
    [12-14-17 4:01:18 pm] | 16024639479 0 /images/Win10-Desktop/d1p2.img ftp://fog:password@10.0.0.23
    [12-14-17 4:01:13 pm] * Deleting remote file: 
    [12-14-17 4:01:13 pm] | Files do not match.
    [12-14-17 4:01:13 pm] | 355111921 0 /images/Win10-Desktop/d1p1.img ftp://fog:password@10.0.0.23
    [12-14-17 4:01:07 pm] * Deleting remote file: 
    [12-14-17 4:01:07 pm] | Files do not match.
    [12-14-17 4:01:07 pm] | 190 0 /images/Win10-Desktop/d1.partitions 
    [12-14-17 4:01:02 pm] * Deleting remote file: 
    [12-14-17 4:01:02 pm] | Files do not match.
    

    Thanks,



  • @coxm said in Image Replication Across Storage Groups:

    Where is it pulling the username/password from and where should it be set?

    In Storage Management, there is a credentials area for every single storage node. Those credentials are the ones that work for the local account that FOG uses for FTP operations. Keep reading: On each storage node/fog server there is a file at /opt/fog/.fogsettings that has a username and password setting. When you update FOG on these, the installer looks to this local file to know what to do. So you must ensure that password and the one in Storage Management are correctly set - otherwise this will just all break the next time you update and you’ll be left wondering why. Here’s an article detailing the .fogsettings file: https://wiki.fogproject.org/wiki/index.php?title=.fogsettings



  • @Wayne-Workman
    Ignore me, I hadnt set the storage node as master in its own group. Now that I have we are getting this.

    [12-14-17 3:14:55 pm] * Type: 2, File: /var/www/fog/lib/fog/fogftp.class.php, Line: 463, Message: ftp_login(): Login incorrect., Host: 10.0.0.23, Username: fog
    [12-14-17 3:14:51 pm] | Image Name: Win10-Desktop
    [12-14-17 3:14:51 pm] * Found Image to transfer to 2 groups
    

    Where is it pulling the username/password from and where should it be set?
    Thanks for your help.



  • @wayne-workman HI, I have made that change to the image file and this is the log output for replication from the master logs

    [12-14-17 3:04:51 pm] | There are no other members to sync to.
    [12-14-17 3:04:51 pm] | Image Name: Win10-Desktop
    [12-14-17 3:04:51 pm] * Not syncing Image between nodes
    [12-14-17 3:04:51 pm] * Attempting to perform Group -> Nodes image replication.
    [12-14-17 3:04:51 pm] | There are no other members to sync to.
    [12-14-17 3:04:51 pm] | Image Name: Win10-Desktop
    [12-14-17 3:04:51 pm] * Not syncing Image between groups
    [12-14-17 3:04:51 pm] | There are no other members to sync to.
    [12-14-17 3:04:51 pm] | Image Name: 
    [12-14-17 3:04:51 pm] * Not syncing Image between nodes
    [12-14-17 3:04:51 pm] | Replicating postinitscripts
    [12-14-17 3:04:51 pm] | There are no other members to sync to.
    [12-14-17 3:04:51 pm] | Image Name: 
    [12-14-17 3:04:51 pm] * Not syncing Image between nodes
    [12-14-17 3:04:51 pm] | Replicating postdownloadscripts
    [12-14-17 3:04:51 pm] * Attempting to perform Group -> Group image replication.
    [12-14-17 3:04:51 pm] * We are node ID: 1. We are node name: POR-FOG
    [12-14-17 3:04:51 pm] * We are group ID: 1. We are group name: Lakeside
    [12-14-17 3:04:51 pm] * Starting Image Replication.
    [12-14-17 2:54:51 pm] | There are no other members to sync to.
    [12-14-17 2:54:51 pm] | Image Name: Win10-Desktop
    [12-14-17 2:54:51 pm] * Not syncing Image between nodes
    

    Looks like it is trying now to do group to group replication but struggling for some reason.



  • @coxm said in Image Replication Across Storage Groups:

    I am guessing that the problem is that the master and slave are in different storage groups?

    That’s not a problem, that’s a choice. and it’s ok because fog can replicate across groups, it’s done in Image Management. Go to image management, click an image, and you should be able to find a sub-area where you can define what storage groups the image belongs to. In this same area, you can set which storage group is the master group for that image.

    When you set this, whichever is designated as the master group is the one where captures will go, and replication goes only from the master group to the non master group. Group-to-Group replication uses the masters in the group. So you need to set GroupB’s storage node as the master for it’s group.

    Also, you should read these fully, I think you’ll learn everything you need from them:



  • Thanks for the reply @Wayne-Workman. I have the master node and the storage node configured and working but the master node wont replicate the imkages over to the storage node. the membership is as follows.

    Fog Server - Master - Storage GroupA - Physical Location A
    StorageNode1 - Not Master - Storage GroupB -Physical Location B

    the logs on the master show the following

    [12-14-17 2:54:51 pm] | There are no other members to sync to.
    [12-14-17 2:54:51 pm] | Image Name: Win10-Desktop
    [12-14-17 2:54:51 pm] * Not syncing Image between nodes
    [12-14-17 2:54:51 pm] * Attempting to perform Group -> Nodes image replication.
    [12-14-17 2:54:51 pm] | There are no other members to sync to.
    [12-14-17 2:54:51 pm] | Image Name: Win10-Desktop
    [12-14-17 2:54:51 pm] * Not syncing Image between groups
    [12-14-17 2:54:51 pm] | There are no other members to sync to.
    [12-14-17 2:54:51 pm] | Image Name: 
    [12-14-17 2:54:51 pm] * Not syncing Image between nodes
    [12-14-17 2:54:50 pm] | Replicating postinitscripts
    [12-14-17 2:54:50 pm] | There are no other members to sync to.
    [12-14-17 2:54:50 pm] | Image Name: 
    [12-14-17 2:54:50 pm] * Not syncing Image between nodes
    [12-14-17 2:54:50 pm] | Replicating postdownloadscripts
    [12-14-17 2:54:50 pm] * Attempting to perform Group -> Group image replication.
    [12-14-17 2:54:50 pm] * We are node ID: 1. We are node name: POR-FOG
    [12-14-17 2:54:50 pm] * We are group ID: 1. We are group name: Lakeside
    [12-14-17 2:54:50 pm] * Starting Image Replication.
    [12-14-17 2:44:50 pm] | There are no other members to sync to.
    

    I am guessing that the problem is that the master and slave are in different storage groups?



  • @coxm said in Image Replication Across Storage Groups:

    I have got 4 office locations and, at the moment, 2 FOG instances. One master at location A and in storage group 1 and one storage node in location B win storage group 2.
    what i want to end up with is a storage group for each location with a storage node in that location (Physically and in FOG config) but the single master in location A. This would then allow image management for all remote storage node locations from my base office location.
    Is that possible?

    Yep. And it’s a good design too. Did you need help with setting it up or just wondering if it can be done or not?


 

318
Online

41.6k
Users

12.1k
Topics

114.4k
Posts