Satellite 6.3.3 RHEL 6 repository selections
I have just built a new Satellite server which has 6.3.3 on it. I went to go to the repos to select and found that the main Red Hat Enterprise Linux 6 Server RPM x64 or i386 repos are no longer listed, but only the minor releases such as RHEL 6.2, 6.3, and so on. This is also the case for child channels too, such as Server 6 Common, etc.
This is from the Satellite server built straight from 6.3.3:
I have a test Satellite system which was built and was running 6.3.2 until I upgraded today. On that system I was able to select the base channel and not have to worry about selecting each minor release. I am using the same Manifest as the test Sat sever. The only difference is that was built on 6.2.2
Here is the same repo list but from a Satellite server built from 6.2.2
Can someone from Red Hat explain why this is happening, a bug?
This looks like only RHEL 6 repos are listed this way.
Responses
Guru
10133 points
Personally, I have the base RHEL 6 and then (for example) while I was transitioning RHEL 6.9 to RHEL 6.10 clients, I had flat rhel 6 and 6.9 and 6.10 channels activated. I found no need to have 6.1, 6.2, 6.3, 6.4, 6.5 and so forth, particularly for my RHEL 6 systems. Once my systems reached 6.10, I stopped ingesting 6.8 from Red Hat from my public satellite, and stopped including it on the disconnected satellites I use. I kept 6.9 and 6.10 *(and I only kept 6.9 because I might need to do a "yum downgrade" of a specific rpm and I want that channel available just in case).
We have one public facing satellite. We do a satellite export of the content view (not channel by channel as the broken documentation states, we use the method described by Red Hatter Rich Jerrido listed here https://access.redhat.com/articles/2390791
After doing the export, I perform the command hardlink -cv /path/to/the/content/view/export because the footprint is 1.7TB and doing this command reduces it from 1.7TB to 270GB or less. Now to preserve this when you send this (for example) to an external hard drive, you need to use
rsync -H -au --progress /path/on/your/satellite/where/you/exported/this/content /your/mounted/large/external/hard/drive
and the -H preserves the links and makes sure when you rsync it, it doesn't expand to it's regular footprint in size. You'll need to do the same when you take it off your external hard drive to go to your disconnected satellite (when I say "disconnected"), I mean it as Red Hat describes it, it's on a network that can not be directly connected to "see" Red Hat's content servers. If you need help with that term, check google for links to Red Hat on that.
The same principles I mentioned for RHEL 6 I follow for RHEL 7. namely, I do not take minor releases (7.1, 7.2, 7.3) when I'm on 7.5. I might keep 7.4 if i need for example to do a "yum downgrade" of a specific rpm (I had to do this last week for just one rpm).
Hope this helps,
RJ
Pro
482 points
This is apparently a bug which looked like it should have been fixed in 6.2.2 but it is showing on my version of 6.3.3. So right now I am stuck and can't do a thing since even some of the child channels show "No repositories accessible. Check that Red Hat Subscriptions have been imported"
Guru
10133 points
Gary,
I may not have seen what context you were asking here. I thought you were asking on guidance on what channels to select for a specific scenario, but if you're encountering this issue, I imagine you've submitted a case for this?
I have not had this issue myself and I'm using 6.2.15 for my public facing satellite. I am able to acquire all the channels I am subscribed to.
Added, you mentioned a bug, would you mind giving the link to the bugzilla you found for others who visit this specific discussion to have clarity on that? I can't tell from that image... Is your entire set of 6 channels grayed out so that you have no ability to select? I ask because I can't definitively determine this from the images. I'm able to select any of the 6 channels I'm subscribed to on 6.2.15. Would you post the bugzilla link?
Regards,
RJ
Pro
482 points
Basically I want to select the base RHEL 6 channel and not have to worry about selecting the minor versions and the upkeep with that when a new version is released. The amount of storage is not an issue since I have the channel syncing on demand. Using the same process for getting the manifest, I used the manifest that is from the Sat 5 system since that has all of my subscriptions attached to it. I changed the version to reflect 6.3 Satellite and downloaded and imported the cert. This was the same process I used when I built a Sat 6 system last year and the same process that I used for my test Sat 6 system that is currently being used.
The second image with the yellow highlight is from the test server which I was able to see the base channel months ago when I built it and selected that to sync, its currently grayed out because it is part of a content view.
I do currently have a case open but it sounds like support is asking me to remove subscriptions from the Sat 5 system and attach them to the Sat 6 system, but I am not ready to import servers to the Sat 6 system yet. So that would leave any clients attached to Sat 5 unsubscribed which could cause issues.
I did try to refresh the manifest, which did correct some channels. At one point I did see the RHEL 6 base channel but then some of the others like RHEL 6 Common was showing as ""No repositories accessible. Check that Red Hat Subscriptions have been imported", even after several refreshes.
Here is the bugzilla I found: https://bugzilla.redhat.com/show_bug.cgi?id=1555468
Guru
10133 points
Gary,
Does this Red Hat Solution https://access.redhat.com/solutions/2099421 fit the issue you describe?
ADDED I can't tell you how many times (sadly) refreshing the manifest has resolved issues https://access.redhat.com/solutions/1527153. It is worth a try.
Guru
10133 points
Hi Gary,
Gary, for your error of "No repositories accessable... - see this link https://access.redhat.com/solutions/3422021, however, I'm curious if that really fits the issue (as a solution) for the problem you are facing.
I have an "overkill" amount of storage, and yet still do the de-duplication, but if you're fine with storage, that's great.
I see it seems you're probably getting ready to migrate your clients from a satellite 5.x server to 6.3.current.
We did that last year - and when we did, we made sure the channels were brought over from our public facing satellite. I suspect you do not have "disconnected" satellites to feed, so I suspect you are just currently dealing with a public-facing Satellite 6.3 server and (just to restate) want to get your channels ingested... but you're getting errors for no repositories.
When I went to the Red Hat Summit in 2014 (and later years too), the Red Hatters there (and my sales guy) said this was the method to transition clients from Satellite 5.x to 6.current... (they said this to hundreds of people, briefed it in presentations) 1) stand up a 6.current satellite with all your subscriptions. The method you described where you flip your current (what used to be called) server certificate to version 6.3 link here https://access.redhat.com/management/distributors/ enables you to take your current footprint of subscriptions and put them (concurrently) on your Satellite 6.3.current server for the purpose of draining from the Satellite 5.x to your new Satellite 6.3.current server. That's the method we used and we drained the clients from the old satellite to the new satellite. not to be master of the obvious here in this paragraph.
Now when I did this, it was understood we could have the same footprint of entitlements on both satellites for the purpose of draining from one to another. That's what we did.
To acquire the channels, what I do from my public facing satellite, ingest my current manifest with EVERYTHING and we have no clients attached. Yes, it's sole purpose in life is to be a means to acquire satellite content from Red Hat, then we carry the exported content view to our "disconnected satellites". Now to acquire the channels from Red Hat (we have this portion in common) I point the cdn url to Red Hat, then make a sync plan of all the channels I want.
I see from your previous reply above, you say...
quote:
Basically I want to select the base RHEL 6 channel and not have
to worry about selecting the minor versions and the upkeep with
that when a new version is released.
In our case, if we have RHEL 6.9/6.10 clients and RHEL 7.4 and 7.5 clients, then we select the corresponding channels to be able to acquire those channels. If our expectation is to have these channels available, then that's the means to acquire the channels (and for us, we have that expectation). We use this method documented by Red hat https://access.redhat.com/solutions/1287373. We synchronize the channels repeatedly (set up a sync plan every hour). I go in and check before an export (which is why we de-duplicate the content view, not applicable for you), and make sure all channels synced properly. In some cases I have to select a specific channel and manually re-sync it. In unique cases, but often enough I have competing tasks I have to kill off with a method given to me by Red Hat because on occasion a task will get in the way. You can see if some task is interfering by going to the tasks in one of the left drop-down menus in the Satellite Web UI.
Let us know how it goes. Red Hat I believe ought to let you take your total footprint of entitlements and concurrently place them on your new satellite for the purpose of draining from one to the other. If you happen to have RHEL 5 clients, that content is still available on satellite 6, and you can synchronize it from Red Hat. We eliminated our RHEL 5 boxes a long while ago.
Let us know how it goes, I or someone will attempt to help
Regards,
RJ
Pro
482 points
Well, I was on the phone with support and shared my desktop with them showing the tech exactly what I was doing; downloading the manifest from the satellite server that has all of my entitlements, importing that into the satellite 6 system, and then showing him that the repos are not listed. After refreshing the manifest, some repos then showed the corresponding channels while another repo came back with "No repositories accessible. Check that Red Hat Subscriptions have been imported" message even though it showed before I did the refresh.
Support said that he "refreshed" the manifest from his side and that I should refresh my manifest in one hour to see if that fixed it, which it did not.
An example is that after a few refreshes of the manifest, the RHEL 6 server now showed all minor releases and the base channel, while the RHEL 6 Supplementary channel (and others) showed the no repositories accessible message. Then on another refresh, the supplementary channel shows everything, but the RHEL 6 server channel shows the no repo message.
I did find that RH article about future dated subscriptions, but I do not have any so that issue would not relate to me.
Again I have used this method on the other Satellite 6 systems, from 6.1.x - 6.3.2 and have never had this issue using the same manifest import method which I was told by the Satellite folks at Summit it would be acceptable to do that.
So again I don't know if this is a bug, if anyone else is having this issue on a fresh 6.3.3 system or what, but I am totally frustrated after many years of dealing with this application that works when it wants to...
Pro
482 points
Support has asked me to create allocate a subscription from the Sat 5 manifest to the new Sat 6.3.3. I used a VDC sub which includes all of the RHEL server channels, removed the manifest from the Sat 6.3.3 system, imported the new one that contained the 1 VDC sub, found that that the same results are happening. Either they are missing their base channel, not showing all minors, or showing that I do not have a subscription for the repo, "No repositories accessible. Check that Red Hat Subscriptions have been imported".
I even went ahead and uninstalled Sat 6, removed all the files from the databases/file systems, reinstalled Sat 6.3.3 and the issue remains, but it varies on which repos are listed.
Surely I can't be the only one with this issue?
I have no idea what else to do........
Guru
10133 points
Gary,
Sorry you're going through this. Where I work we stayed at 6.2.15 (still supported) for production requirements.
Regardless, I'll ask some others I'm associated with to look at this discussion and research some as well.
RJ
Guru
10133 points
Gary,
With your recent satellite reinstall, when you ran the satellite-installer, did it give any negative or failed actions in the logs?
Anything from a foreman-debug or foreman-debug?
Anything useful from (typing from memory) /var/log/production or /var/log/foreman*/* logs?
I imagine you've given logs to Red Hat.
Guru
10133 points
I've asked some others in the Red Hat Accelerator Program to look at this to see if any of them faced this as well. It may take a bit for them to reply.
I'd recommend to keep pressing on with support, and if it is warranted, elevate (if it is warranted to do so).
Guru
10133 points
Gary, another person in the RH Accelerator Program is looking at this, I believe he's in the same boat. Keep pressing support, and if warranted, elevate (if actually warranted)
Regards,
RJ
Red Hat
Community Member
60 points
Hello RJ and Gary, good afternoon.
Case updated. Please let us know if helps you.
Just waiting for your reply. :-)
Have a good one folks. Waldirio
Guru
10133 points
Gary, I had a friend in the Red Hat Accelerator group who said he had a similar issue, but he said:
The weirdness has resolved itself in the new Org with the new manifest.
I did sync a new product in at 7.5 and after several hours all of the repos
appear, including 6Server.
Not sure that tidbit is useful, but work with Waldirio from Red Hat
Regards,
RJ
Gary,
I had the same issue with mine. For me I just had to make sure Location was set to any-location and then I was able to see the base RHEL 6 repo. I've noticed the same thing when dealing with Manifests.
Hey,
I had the same issue with mine. Satellite 6.3.4.
Pro
482 points
When I was working with support they said it was a bug and that a bugzilla would be opened for it. But I do not have that number. Here is the fix that I had applied:
Make the following change under the "production" section in /etc/foreman/database.yml: production: adapter: postgresql database: foreman username: foreman ... pool: 20 <-- change from 5 to 20 Save changes and restart Satellite services: $ katello-service restart Then double-check the available repositories.
Welcome! Check out the Getting Started with Red Hat page for quick tours and guides for common tasks.