Hi guys,
after I reboot/restart my vCenter appliance (virtual machine) vmware-vpxd won't start and fail within initialization (everytime). After 3 days of researching I have a workaround:
1. ssh into vCenter appliance and run command: ps -ef | grep vpxd
2. here I can see, that there is something locked out - connected with vmware-vpxd service (flock -on ....)
3. run command service-control --stop --all
4. then I kill all processes (with command kill), which are listed as output from command ps -ef | grep vpxd (connected with vmware-vpxd service)
5. run command service-control --start --all
6. when services are starting I have to type password, which is stored in (/etc/vmware-vpx/embedded_db.cfg) - postgres database password
After this process everything is fine and working, but I don't have an idea how to fix this problem permament. I don't want to do this workaround after every restart. My vCenter version is 6.0.0U3f and my idea was to upgrade to version 6.5.0 (but actually not possible). Any idea, how to fix this problem without migrating to 6.5.0?
Thank you for your answer.
Jakub
The first step is to look at (and/or post) the log file to see what's actually being logged. If you have the ability to open an SR with VMware, you may want to get one of those started as well.
Dear daphnissov,
yeah of course. I will look up into logs and I'll post there. Should I post vpxd.log and what else?
Jakub
I'd recommend you look at your own logs first (start with vpxa.log) and see what you can find. Use some Google skills to see if there are known KBs, and search the communities here. One should make every attempt to help themselves (and potentially learn a great deal in the process) before asking for help.
Dear daphnissov,
I understand and I took maybe 3-4 days (google, known KBs, the communities here) and I don't find any the solution for my problem, unfortunately. Then I opened vmware-vpxd script in text editor, debug code and find out, that there can be a problem with locking. I think, that can be a very specific strange problem, so for that reason maybe the support of vmware is the best way, right?
Jakub
But you haven't even looked at the log yet, so why are you even doing that? Look at the log first at the entries created after you attempt to start the process and see what it says.
Sorry for misunderstandings, of course I looked at the logs, and so many times , but I can't find anything, which can solve my problem. It's bunch of logs and a lot of dependencies and sometimes it can be specific and so hard to find out, where is a problem.
Jakub
I'd recommend restarting your vCSA, then attaching the vpx log to a response. Please do not copy/paste raw text into your response.
Hi guys,
after one month issue occurs again. Below I attached screenshots.
As you can see there is a problem with a flock -on ..., which I mentioned in above post.
As you can see there, the process is waiting for the DB to start up, but I waited and nothing happend.
There is vmware-vpxd.log, in vpxd.log there is nothing logged about this problem, only logs before restart VMs, when it was working correctly.
Then I used my workaround mentioned in my first post and everything was ok.
Any ideas?
Jakub
Can you check if enough is available for the mount points.
# df -Th
Of course it is, as I mentioned in my first post, when I manually stop all services, kill the process manually, start all services and type password for database everything is fine. Maybe (I don't know why), when it is starting automatically after restart, the process want a database password, but nobody doesn't provide that and process locked out? I don't have any idea, where can be a problem and I was searching for any solution, but without success.
Thank you in advance.
Jakub