After patching some virtual systems become unresponsive
After a scheduled patching some systems will be unavailable via RDP or vSphere. Using vSphere we will try to initiate a guest restart and that will not work. As a last resort we have found that we have to just power off the system. After the system shuts down we will power it back up and it will finish patching. Has anyone else experienced this behavior?
1 Comment
[ + ] Show comment
-
Jason thanks for the response. I know I have worked with you and other techs on this issue but no one really has an answer for it besides doing enabling debug and then doing a capture. I recently redid our QA environment and last month I had 5 servers that showed this behavior. Nothing special was done to them and then this month only one server (that was not one of the 5 from the month before) had this happen. I can't predict which servers will have this issue so randomly picking a server to capture is very hard. - Joe Cain 9 years ago
Answers (1)
Please log in to answer
Posted by:
jknox
9 years ago
There are a lot of variables that aren't accounted for, such as the operating system(s) affected, what K1000 and agent version, etc.
I'd suggest enabling debug on one of the affected clients, replicating the patching scenario and getting a KaptureState. Also gather a screenshot of the hardware configuration of the VM, showing cpu sockets, memory allotted, etc.
Once you have that, open a support ticket.
I'd suggest enabling debug on one of the affected clients, replicating the patching scenario and getting a KaptureState. Also gather a screenshot of the hardware configuration of the VM, showing cpu sockets, memory allotted, etc.
Once you have that, open a support ticket.
Comments:
-
Jason thanks for the response. I know I have worked with you and other techs on this issue but no one really has an answer for it besides doing enabling debug and then doing a capture. I recently redid our QA environment and last month I had 5 servers that showed this behavior. Nothing special was done to them and then this month only one server (that was not one of the 5 from the month before) had this happen. I can't predict which servers will have this issue so randomly picking a server to capture is very hard. - Joe Cain 9 years ago
-
I understand your frustration, but unless Support has detailed logs from the clients and/or are able to duplicate the issue, it's difficult if not impossible to track this sort of thing down.
Enabling debug should not add noticeable overhead, so it shouldn't hurt anything to leave it enabled on servers. It basically just makes the logging that the agent already does more verbose.
That said, I would enable it on one test server first to be absolutely certain. - jknox 9 years ago
