27 มีนาคม 2555 20:58
I had a 2 server array that hung when activating the configuration. When this happened, I let it run for 24 hours and then I killed the console and the UAG configuration went corrupt. After that I tried to import a previous version, but struggled for 3 days to get everything (ie NLB) working correctly again.
I gave up and t rebuilt the array from scratch. Guess what… its happening again. Right now im just letting it run, but it may be too late. Has anybody experienced this before? If so do you have any ideas how to kill the activation without corrupting the UAG configuration?
F.Y.I, I just ran a command line synchronization (configmgrutil -a ) and it says that it was successful, but the GUI is still running. I’m going to let it run and if someone has some suggestions for me to save my config, I’d really appreciate it.
28 มีนาคม 2555 13:23I don't know of a way to pull the configuration from the file system (it's encrypted), but regarding your core problem of why this is happening, do you have a firewall between the internal NICs of the UAG servers and/or between the UAG servers and the corporate network? If you do, try temporarily allowing all traffic between the two UAG nodes and between the UAG nodes and the internal network and see if that makes a difference. Who knows, doing so might even allow your currently running activation to suddenly finish. I have seen internal firewalls present various problems with activation in the past.
28 มีนาคม 2555 14:17
Thanks for the response. There are no firewalls between the two uag, they are on the same vlan. The only thing between uag and internal network is they are on different vlans (on the same switch). Not sure if it's worth mentioning, but they are a vsphere 5 guest vm's.
I just checked and its still giong from yesterday.
28 มีนาคม 2555 17:35
Another observation is, after a while I noticed that the activation window "activate" button becomes un-greyed and you can click it again and becomes grey again. Also the activate dialog has shadows on it, like a video issue or software bug.
After reading this:
I tested DNS and name resolution seems to be working fine and then I re-ran configmgrutil -a (with ouput and debug set to 1) and I noticed that it zips along and then hangs for a few seconds on the below, then zips to the end:
m_msg_type = [CLUSTER_MSG_TYPE_DEBUG]
m_msg = [The file [C:\Program Files\Microsoft F refront Unified Access Gateway\von\conf\WFLB.xml] was written successfully.]
m_cluster_node = [myservername01]
I tried restarting SQL and TMG services; I also tried an iisreset without change. At this point I shut down all the TMG, then UAG services, then hesitantly killed configuration.exe in task manager and rebooted the array manager server. After reboot, I opened the array monitor and everything converged and synced without errors. I re-ran configmgrutil –a without errors and it appears that I don’t have any corruption (knock on wood).
I suspect the issue has something to do with my federated trunk and ADFS somehow and maybe some kind of software bug with TMG or UAG. So from here on out, I’m not going to use the GUI for activation and I’m going to stick with the CLI for
activating. I also turned on “informational messages”, as outlined in the above link. So maybe if it happens again, I’ll be able to tell where the issue happens. I think another precaution I’m going to do is take VMware snapshots prior to activation of a new configuration.
In the meantime, if anyone is experiencing this issue and/or maybe has a fix, please let me know.
- แก้ไขโดย Bret Gardner 28 มีนาคม 2555 17:45
5 เมษายน 2555 15:51
Here is a follow-up post. As previously mentioned, I was going to only use the command line interface, but it appears that the CLI activation is limited in some way and I don’t see my changes after running it. Therefore I went against what I said and have been using the GUI.
After a week, it finally happened again. As previously mentioned, I have "information" turned on and it indicates that activation has completed successfully. Notice that the activate button is no longer grey in the snip; so by pressing the button again, it appears activate again and the dialog box stays. So I think as long as it says activation complete, I feel that it is safe to kill the process. I think that when I previously killed the process, the activation must have been running, which killed my UAG install. If others are getting this and it is bug in UAG, please mark as answer, so maybe ms will fix it.
- ทำเครื่องหมายเป็นคำตอบโดย Bret Gardner 3 มกราคม 2556 23:48