Some Subscriptions not visible to guests

Latest response

I am having some trouble with a subscription not being available to guests on a hypervisor.

I have multiple hypervisors running the full RHEL and running virt-who on each. The subscription for Linux running on each of these is the Linux with Smart Virtualization and Management. In addition there are three subscriptions for "Extended Lifecycle Support Unlimited Guests."

On one hypervisor, it works as expected. both subscriptions are listed for the hypervisor, and the guests needing the ELS subscription can have both added to them.

On two different hypervisors, the situation it different. Although the hypervisor lists both subscriptions, the guests can only see the main Linux with Smart Virtualization subscription. When I try to add the ELS guest subscription, it is not listed as one that can be added.

I have checked the rhsm.log and all three hypervisors list adding the subscription and give a pem certificate name. Using rct cat-cert <certificate.pem> shows the name and subscription info for the ELS subscription.

I have restarted virt-who several times, and have removed and re-added the subscription to they hypervisors, but so far, nothing seems to work. The guests on two of the hypervisors still cannot see the ELS guest subscription.

Anyone have any ideas which logs I can look at to see why they are not available to the guests?

Thanx in advance.

Responses

Who knew that I would be dragged, kicking and screaming, back into Windows trouble shooting? In frustration over memory usage, I rebooted the box. After the reboot, the subscriptions magically appeared.

The 3 R's of Satellite 6/Windows trouble shooting - 1) Retry, 2) Reboot, 3) Re-install.

Frank - Thanks for reporting back with the message that the guest VMs are now seeing the expected subscriptions. You're right, though, in that a reboot of the Satellite server should not have been necessary to get this working. Just checking... was it the Satellite server you rebooted?

Just for the record, can you please confirm the version numbers for:

  • Red Hat Enterprise Linux
  • Red Hat Satellite
  • virt-who (package)

I'm just curious if any of these might have contributed to the issue.

Off-topic, but as a former Microsoft Windows Server administrator, having begun in Windows Server NT, I feel your pain.

Russell - Yes, it was the Satellite server that was rebooted. I was frustrated over its use of the swap file and the slowness that accompanies it.

For the record, here are the versions you asked for.

  • Red Hat Enterprise Linux
    • Sat Server - 7.3
    • Hypervisor - 7.3
  • Satellite - 6.2.10
  • virt-who - virt-who-0.19-2.el7sat.noarch

Frank - Thank you. From what I can tell, you are using the very latest version of all three components involved. I wanted to check, in case you might have had a known issue occur. I haven't heard of anything similar occurring to other customers.

Since the symptoms were resolved with the reboot, there's not much likelihood of finding the root cause. However, you were concerned about the usage of the swap file, which suggests memory usage at the time was outside what you would consider normal. I wonder which processes were using the most amount of memory, and why?

Russell - Memory on that box is ALWAYS outside of normal. Since Satellite 6 is a chimera - many different packages crammed into a single entity without any real unity - it is extremely wasteful of resources. Pieces of it are programmed in Ruby and Python with a touch of Java thrown in for good measure. But where memory suffers is in the fact that it runs both Postgresql and MongoDB on the same box at the same time. All databases are greedy when it comes to memory, but having two on the same box at the same time assures that the only time there is any free memory is just after a reboot.

My problem is that I believed the prerequisites listed for Satellite 6.0 when I first tried to install it. The memory requirements at that time were 8 G of memory and 4G of swap, but they recommended 12 G of memory. I happened to have a box that was free that had a large local disk and had 12 G of memory, The requirements of 6.2 are now 12G minimum with a recommended 16G. I am currently stuck with 12G of memory and 6G of swap space. At least once a month I get warning that I am using 94% of my swap space.

Since purchasing a new box is not an option at this time, I end up usually rebooting the box once a month. Although lately, I get the warnings more an more often, so I am thinking of moving it to an automatic reboot once a week.

Lesson learned. If the prerequisites are for a large amount of memory, check to see how many different database engines will be running. If there are two, quadruple the memory. And add more just for good measure.

Frank - Thanks for your reply. For all the reasons you mentioned, memory usage can be a problem. Unfortunately, the memory requirements of Satellite 6 have increased, as you already mentioned. That's left yourself, and likely other customers, in an awkward position because it's not a simple matter to upgrade the memory in a physical host.

I can't provide any immediate tips about controlling memory usage. Since this is proving to be a problem, though, I suggest you raise a support request with Red Hat Technical Support.

Close

Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.