• Hello

    I have 15 locations through the state. They are connected by MPLS Lan. I have about 15 machines a couple of locations and was wondering what is the best way to image them. What are some best practices of working with a subnet. Or is it a case I should pickup the machine and bring it back to image. Or can I can image them in place… Can you pixe boot across the network. Any suggestions and help with appreciated.


  • Moderator

    @cnkpadobi FOG Settings -> FOG Linux Service Sleep settings

    Not sure if there’s an easy way to track if replication is finished, to be honest. I don’t use multiple nodes myself, so not an expert on replication specifically.

  • @Quazz Well wanted to jump start getting that image over to the remote location so my staff can start imaging without waiting for the long replication time to finish over the 1.5mpls. I only want to move 1 image over to that storage mode.

    Thanks I will check that.

    How I change the service checks…

    Is there a way to tell if the replication completed or what percentage it is at?

  • Moderator

    @cnkpadobi Apache/php may be using the wrong timezone.

    As for the replication, only other way would be to manually copy the image folders to external storage and then attach that to the other machine when it’s done and copy it back to their appropriate locations. Should work, but I’m not sure why you’d want to do this.

    You can change how often the service checks in in FOG settings instead of using cron jobs, by the way.

  • @george1421 I ran a replication overnight and was wondering is there a way to know for sure if the job completed. When I look at the log file I see the following. It hard to tell but the time on the log is off…

    [06-20-17 2:34:51 pm] | Replication already running with PID: 65438
    [06-20-17 2:34:51 pm] | Image Name: FWDTI3BASE
    [06-20-17 2:34:51 pm] * Found Image to transfer to 1 node
    [06-20-17 2:34:51 pm] * Attempting to perform Group -> Nodes image replication.
    [06-20-17 2:34:51 pm] | There are no other members to sync to.
    [06-20-17 2:34:51 pm] | Image Name: FWDTI3BASE
    [06-20-17 2:34:51 pm] * Not syncing Image between groups
    [06-20-17 2:34:51 pm] | dev/postinitscripts: No need to sync fog.postinit file to FWTC
    [06-20-17 2:34:27 pm] | File Name: dev/postinitscripts
    [06-20-17 2:34:27 pm] * Found Image to transfer to 1 node
    [06-20-17 2:34:27 pm] | Replicating postinitscripts
    [06-20-17 2:34:27 pm] * Started sync for Image postdownloadscripts
    lftp -e ‘set xfer:log 1; set xfer:log-file “/opt/fog/log/fogreplicator…transfer.FWTC.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/postdownloadscripts” “/images/postdownloadscripts”; exit’ -u fog,[Protected]
    [06-20-17 2:34:27 pm] | CMD:
    [06-20-17 2:34:27 pm] * Starting Sync Actions
    [06-20-17 2:34:23 pm] * Deleting remote file: /images/postdownloadscripts/
    [06-20-17 2:34:23 pm] | Files do not match.
    [06-20-17 2:34:23 pm] | 233 0 /images/postdownloadscripts/fog.postdownload ftp://fog:06308b@
    [06-20-17 2:34:10 pm] | File Name: postdownloadscripts
    [06-20-17 2:34:10 pm] * Found Image to transfer to 1 node

    My last question is… Since this location is physically close is there another way to get the image to the storage node without replicating it over the network

  • Moderator

    @cnkpadobi Not sure I’m following…

    But the name of the service is FOGImageReplicator

    You could create a bash script to stop and start the service or call your service manager directly from cron. You will probably edit the /etc/crontab file to set this up.

  • @george1421 Thanks George… another question,

    If I set a cronjob up in linux to start and stop the replication. Which process do I used to do this.

  • @george1421 @Sebastian-Roth The issue with pxe boot has been resolved.

  • Moderator

    @cnkpadobi The picture you posted is showing “No boot filename received”. This sounds like the DHCP in this subnet is handing out IPs but not nextserver and filename information for PXE booting. What kind of DHCP server is serving this subnet and how are option 66 and 67 configured?

  • Moderator

    @cnkpadobi Ok at the remote site, what is your dhcp server?

    What do you have configured for dhcp option 66 and 67?

    Do you have a fog server at that location or only HQ?

    By the way the picture explains more than just the error message. Well done!

  • @george1421

    When trying to pxe from subnet I get this message… it work fine from headquaters!0_1497641254213_2017061695114526.jpg

  • Moderator

    @cnkpadobi There should not be anything to keep the fog server from responding to those other subnets. Just make sure you have the gateway set correctly on the fog server and it should route to the other subnets without issue.

  • OK thanks, I think this is my last question … do I need to configure fogserver to accept ip address from my other subnet.

    for example HQ is 192.168.0.x
    subsite is 10.10.10.x

  • Moderator

    @cnkpadobi This question is a bit complex to answer, but lets take the bullet items first.

    1. There is not a way built into fog to do this. It would be a nice feature to add. With that said you can do something similar in linux. Use a cron job (i.e. scheduled task in ms windows) to start and stop the fog image replicator task. This way you can have the replicator run after hours and not during business hours. You can also adjust the replication bandwidth based on time of day with similar trickery.
    2. Yes, each image has a check box which enables it for replication or not. By default this box is enabled. Turn it off and the image is excluded from replication to all storage nodes in all storage groups.
    3. A storage node requires slightly less than a full fog server. With that said I’ve run a full fog server on a Raspberry Pi3. For 1 or 2 unicast images that Pi is working hard but able to keep up with image deployment. The system requirements for a FOG server or storage node are not much.
    4. Well this one is a bit hard to answer. Will it work in hyper-v yes. Will it work in vmplayer yes, but I wouldn’t use it for a production environment (I think you will break the VMWare eula if you use vmware player for this anyway). If running on a type-2 hypervisor you will also be limited to the throughput of your host os and hardware. Its a complicated subject to explain. But given the choice of a circa 2009 desktop or running FOG on a type-2 hypervisor, I would pick the 2009 desktop.

  • Thanks, guys, If I turn off replication and do not want to craputre via the remote subnet. Then I would just copy the preexisting image to the storage node prior to imaging?

    I have a few questions,

    1. Is there a way to set the time interval when replication could occur.
    2. Is there a way to limit which image sync.
    3. Is there a specific hardware specs for the Storage node
    4. Is it ok to run on VMWARE Player or HYper-V… If so is there any pros or cons?

  • If the connections are to thin for gigabytes i would suggest to work with local storage nodes without replication if not mandatory necessary.

  • Moderator

    So what’s your MPLS built on T1/OC3? That will dictate the bandwidth between the sites.

    If you are using a T1 then probably you will want to setup a storage node at each location then you can image locally. This storage node is almost a full FOG server but it doesn’t have a database. The storage nodes use the Master FOG server’s database. Just realize in this setup if you need to capture images at the remote location a storage node solution isn’t probably the best choice. But if you develop your golden images at HQ and deploy at the sites it will work pretty good.

    FOG can image across subnets no problem. if your network is setup correctly. You just need to pay attention to the subnet to subnet transfer rates. You have to remember you will either be replicating multi-GiB images or imaging multi-GiB images. Your link speed will need to be taken into consideration.