Active multicast tasks/groups, etc.
I’ve got a couple of lingering issues in 1.1.0 that someone can (hopefully) help me sort out.
As I’ve mentioned in a previous thread, if I reboot my FOG server I’m being prompted twice within a couple of minutes to update the database schema. Restarting the mysql service clears this up and lets me log in, but I’m hoping this is not an indication of a bigger issue.
Under task management for active snapin tasks and active tasks, the start times for jobs seem to change randomly (always adding 5 hours). Sometimes they will look correct and then other times I can log in from another computer (or even the same computer) and the start time will be listed as 5 hours later than my time zone.
Also, I’m currently fighting through getting multicasting to work reliably so that our techs can reimage 2000+ computers this summer. I don’t recall having this problem with 0.32 but in initial testing with 1.1.0 I can create a group and create a multicast task. The number of hosts may look corrrect before I actually deploy the task, but when we get (say, in this case 30) computers PXE booted and the task doesn’t start…I can look in “Active Multicast Tasks” and instead of 30 clients there may be double that number. In another instance today, I created a multicast task under a particular group (of 3 computers) and the image deploy went fine - but while that task was working I created another multicast task of 2 computers under a different group. When I searched for the multicast task created for the 2 computer job, the number of clients in the “Active Multicast Tasks” had actually increased from 3 to 5 in the job that was in the middle of deploying - as if the 2 were added to the currently running image. When the 3 computer multicast task was finished, it cleared out everything as if I never created a new multicast task under with a different group.
Any insight would be greatly appreciated,
Ubuntu 13.10 Server
Please test 1808,
I believe I have narrowed and corrected for this issue.
Thank you all,
I know there’s an issue and am working on it currently. I even think I know WHY it’s happening, but don’t know how to fix yet, working…
1807 has given me no further reliability in multicasting - I’m still seeing issues like new group multicast tasks getting dumped into already running multicast tasks. Were you referring to installing MySQL with no root password as a solution to multicasting problems or just the repeated prompts to upgrade the DB schema?
Thanks for the reply. No apology necessary on MySQL - it’s not a showstopper by any means and I know how much work you’ve put into this project and for that I’m grateful. I will update to 1806 and see how it goes.
Yes, that was me who built the ipxe.pxe file with success for 1.0.0.
I’m hoping I can get 1.1.0 running reliably enough to roll out in our school district within the next 2 weeks. I want to thank you and the other devs for the progress made from 1.0.0 to 1.1.0. The iPXE/undionly.kpxe combination has been working well and I’m blown away by how almost all of our models in production (14+ Dells and 5 HPs) have been working with the stock bzImage kernel with no arguments necessary. I will update the working model thread when I get through them all.
I don’t know what the database issue is and for that I’m sorry. I haven’t been able to replicate it no matter what I do. From what I hear though, if you make the mysql user password blank or none, it doesn’t crash. I don’t know why, but it’s just ramblings from my research. It may not be accurate.
I found a few issues with multicast and hopefully have fixed them in SVN 1806 if you’re willing to give it a shot.
Also, I believe, if my memory serves correct, you were one of the guys with the net1/net0 issue and you built your own ipxe.pxe file based on the default.ipxe with net1/mac sent vs. net0. Hopefully 1806 fixes this as well. Basically I’m sending up to three mac addresses at one time to find the host. If the host is present in any or all of those mac’s sent, it should return. Please test and let me know if it’s working for you.