Repo update for 5.2.18? (#17950)

[tlhackque]

---

ModEdit; related ticket: #17950: 5.2.18 still not in fedora 28 repository

---

The dnf (yum) repository is still announcing 5.2.16, although 5.2.18 was released a couple of days ago.

My host is newly updated to fedora 28, so this is the first time in a long time that I haven't had to build from source!

Is the repository usually updated at the time of announcement? If not, what's the expected latency?

[Martin]

The version 5.2.18 still isn't announced/listed in the redirected repository for F28 even when the files are there for F26.

Index of /virtualbox/rpm/fedora/26/x86_64
  Name                                                  Last modified      Size
  Parent Directory
  repodata/                                             14-Aug-2018 18:14  -
  VirtualBox-5.1-5.1.28_117968_fedora26-1.x86_64.rpm    14-Sep-2017 14:23  69M
  VirtualBox-5.1-5.1.30_118389_fedora26-1.x86_64.rpm    16-Oct-2017 14:12  69M
  VirtualBox-5.1-5.1.32_120294_fedora26-1.x86_64.rpm    16-Jan-2018 10:28  69M
  VirtualBox-5.1-5.1.34_121010_fedora26-1.x86_64.rpm    26-Feb-2018 21:06  69M
  VirtualBox-5.1-5.1.36_122089_fedora26-1.x86_64.rpm    16-Apr-2018  9:49  69M
  VirtualBox-5.1-5.1.38_122592_fedora26-1.x86_64.rpm    09-May-2018 18:33  69M
  VirtualBox-5.2-5.2.0_118431_fedora26-1.x86_64.rpm     18-Oct-2017 11:28  71M
  VirtualBox-5.2-5.2.10_122088_fedora26-1.x86_64.rpm    16-Apr-2018  9:48  69M
  VirtualBox-5.2-5.2.12_122591_fedora26-1.x86_64.rpm    09-May-2018 18:32  69M
  VirtualBox-5.2-5.2.14_123301_fedora26-1.x86_64.rpm    29-Jun-2018 20:15  69M
  VirtualBox-5.2-5.2.16_123759_fedora26-1.x86_64.rpm    17-Jul-2018  9:37  69M
  VirtualBox-5.2-5.2.2_119230_fedora26-1.x86_64.rpm     23-Nov-2017 12:06  71M
  VirtualBox-5.2-5.2.4_119785_fedora26-1.x86_64.rpm     19-Dec-2017 13:38  70M
  VirtualBox-5.2-5.2.6_120293_fedora26-1.x86_64.rpm     16-Jan-2018 10:26  69M
  VirtualBox-5.2-5.2.8_121009_fedora26-1.x86_64.rpm     26-Feb-2018 21:05  69M

Index of /virtualbox/rpm/fedora/26/x86_64
  Name                                                  Last modified      Size
  Parent Directory
  repodata/                                             14-Aug-2018 18:14  -
  VirtualBox-5.1-5.1.28_117968_fedora26-1.x86_64.rpm    14-Sep-2017 14:23  69M
  VirtualBox-5.1-5.1.30_118389_fedora26-1.x86_64.rpm    16-Oct-2017 14:12  69M
  VirtualBox-5.1-5.1.32_120294_fedora26-1.x86_64.rpm    16-Jan-2018 10:28  69M
  VirtualBox-5.1-5.1.34_121010_fedora26-1.x86_64.rpm    26-Feb-2018 21:06  69M
  VirtualBox-5.1-5.1.36_122089_fedora26-1.x86_64.rpm    16-Apr-2018  9:49  69M
  VirtualBox-5.1-5.1.38_122592_fedora26-1.x86_64.rpm    09-May-2018 18:33  69M
  VirtualBox-5.2-5.2.0_118431_fedora26-1.x86_64.rpm     18-Oct-2017 11:28  71M
  VirtualBox-5.2-5.2.10_122088_fedora26-1.x86_64.rpm    16-Apr-2018  9:48  69M
  VirtualBox-5.2-5.2.12_122591_fedora26-1.x86_64.rpm    09-May-2018 18:32  69M
  VirtualBox-5.2-5.2.14_123301_fedora26-1.x86_64.rpm    29-Jun-2018 20:15  69M
  VirtualBox-5.2-5.2.16_123759_fedora26-1.x86_64.rpm    17-Jul-2018  9:37  69M
  VirtualBox-5.2-5.2.18_124319_fedora26-1.x86_64.rpm    14-Aug-2018 18:07  69M
  VirtualBox-5.2-5.2.2_119230_fedora26-1.x86_64.rpm     23-Nov-2017 12:06  71M
  VirtualBox-5.2-5.2.4_119785_fedora26-1.x86_64.rpm     19-Dec-2017 13:38  70M
  VirtualBox-5.2-5.2.6_120293_fedora26-1.x86_64.rpm     16-Jan-2018 10:26  69M
  VirtualBox-5.2-5.2.8_121009_fedora26-1.x86_64.rpm     26-Feb-2018 21:05  69M

Can someone from Oracle please check what is happening here?

[michel-slm]

5.2.18 is still not available today. I checked the repodata files and 5.2.18 seems to be announced correctly in repodata/3c2d03747408fb4deb9ace92906c291f156a0ca68711440baf64a6366f61d38a-primary.xml.gz -- but I wonder if some checksums are incorrect

[tlhackque]

Good to know I"m not the only one who noticed

I created a bugtracker ticket for this, referencing this topic. https://www.virtualbox.org/ticket/17950

Hopefully that's monitored more frequently than this forum.

[socratis]

Hopefully that's monitored more frequently than this forum.

The opposite actually. There are a lot more eyes in the forums, a lot more people/volunteers that are willing to test your claim/issue. If it's confirmed, then you go to the bug tracker. For example, just the past week there were 5 similar tickets like yours that were closed as Fixed/Invalid. What they've accomplished (without first getting a confirmation/2nd opinion) was to waste a developer's time.

Always come to the forums first. And then IRC. The bug tracker should be the last resort...

I updated your first post with the ticket information.

[tlhackque]

The opposite actually. There are a lot more eyes in the forums, a lot more people/volunteers that are willing to test your claim/issue. If it's confirmed, then you go to the bug tracker. For example, just the past week there were 5 similar tickets like yours that were closed as Fixed/Invalid. What they've accomplished (without first getting a confirmation/2nd opinion) was to waste a developer's time.

Always come to the forums first. And then IRC. The bug tracker should be the last resort...

In this case, I disagree. You might want to read the whole thread carefully.

You will find that indeed, I came here first. 5 days ago. After waiting several days for the repo to be updated. In the meantime, two other people confirmed that it's still an issue. One pleaded for "someone from Oracle to investigate" a day and a half before I decided it was time to put the issue in a forum that might be monitored more frequently.

I don't think it's unreasonable to create a formal ticket for an incomplete release 8 days after a release announcement. I probably should have done it sooner - this is not something that a volunteer can fix.

Nor is it a waste of developers' time to finish making their work accessible to their users.

If in fact there were 5 other tickets on this issue that were closed as "invalid", something is very broken. The directory listings posted by OTHERS in this thread confirm my report (and match the listings that I [still] get here).

If your comment is an automatic response to another ticket - that issue can be discussed there. But again, I would urge you to read carefully before replying. Your comment was that under light load, you have not experienced that problem -- but I explicitly stated that those were the conditions of my experience.

For your information: I have spent many years as a developer of complex software, and I don't consider my time wasted when additional data - or even duplicate data - is presented. It may help- me to identify a problem. Or at least to prioritize it.

[socratis]

My comment was indeed a paraphrased copy/paste comment that I usually post on the bug tracker. Original:

It's usually better and faster, if issues get first addressed in the VirtualBox forums, a lot more eyes. More than 95% of the issues are resolved over there, which keeps the developers focusing on the bug fixes and enhancements, and there is no need for another ticket to keep track of. For example, yours is most probably not a bug and someone from the developers has to deal with it and close it as "Invalid".

So, if you can, please open a new thread in the <appropriate_link> section of the forums. Please be sure to mention that you came from the bug tracker and include the ticket number.

That comment alone, has resolved hundreds of tickets. And with a success rate of 90-95%.

Do you know how many tickets have been opened because the OP didn't:

• enable VT-x in their BIOS?
• enable the USB controller and couldn't get USB passthrough?
• enable 3D acceleration and complained about slow performance?
• update their ExtPack and there was a version mismatch?
• update the guest additions that came with their default Linux distro?
• ...
• do a gazillion of things, that, if they had opened a thread here instead of the bug tracker, the forum users would have directed them accordingly?

Regarding this specific issue, as I had already replied to you in the ticket, the glitch with the symbolic links is a known one, not from day 1, but from hour 1. There was/is a kink in the server and some links, although valid, do not show up. They're working on trying to figure out what in the seven kingdoms is wrong. And how to avoid such a hiccup in the future.

I don't think it's unreasonable to create a formal ticket for an incomplete release 8 days after a release announcement.

Not at all. Absolutely, 100% with you...

Nor is it a waste of developers' time to finish making their work accessible to their users. If in fact there were 5 other tickets on this issue that were closed as "invalid", something is very broken.

Having 6 tickets open, instead of 1 centralized (where other people would confirm the behavior), is a kind of lazy behavior from the users' part, you got to give me that. If there was one ticket, I wouldn't have anything to say. But... 6? On the same subject? In two days? Ain't that a little bit excessive? Plus, you can always re-open a ticket that's been closed, no questions asked.

Your comment was that under light load, you have not experienced that problem

I did in fact installed a F28 VM and I tried it. And that's why I said, download it manually until it gets resolved. That's a workaround until a permanent fix is established. That does not mean that a ticket shouldn't have been opened, just that there is a workaround.

[michel-slm]

Hopefully that's monitored more frequently than this forum.

The opposite actually. There are a lot more eyes in the forums, a lot more people/volunteers that are willing to test your claim/issue. If it's confirmed, then you go to the bug tracker. For example, just the past week there were 5 similar tickets like yours that were closed as Fixed/Invalid. What they've accomplished (without first getting a confirmation/2nd opinion) was to waste a developer's time.

Always come to the forums first. And then IRC. The bug tracker should be the last resort...

I updated your first post with the ticket information.

TIL there's an IRC channel!

For what it's worth, I think the reason this might have slipped through the developers' attention is the subject does not clearly state which repo is affected (in this case, yum). Could someone edit?

[socratis]

"TIL" = Today I Learned (???)
If that's the case, have you taken a look at the Community section? It's one of the few perma-links on the left side of the site.

I think the reason this might have slipped through the developers' attention

It hasn't. Have you taken a look at the "Discuss the 5.2.18 release" (2nd post)? Yeah, I'd say that something slipped from someone's attention...

Could someone edit?

It affects more than one repo. And it's not as easy as it may sound, when a symlink is there, a symlink shows up, but when you're trying to download it, it's not. And there's so much you can do with a web/download server that's not in your control.

Add the summer vacation downtime, and you're in for a perfect storm..

[michel-slm]

Have you taken a look at the "Discuss the 5.2.18 release" (2nd post)? Yeah, I'd say that something slipped from someone's attention...

Missed that thread, thanks!

[srondeau]

I ran into the same issue, and resolved it with a special invocation of dnf to directly access the release 26 repository instead of going through the release 28 symlink to the release 26 repository:

dnf --releasever=26 -y update VirtualBox-5.2

Edit -- disabling the modular repositories may also be necessary:

dnf --releasever=26 --disablerepo=fedora-modular --disablerepo=updates-modular -y update VirtualBox-5.2