Life without FreeBSD Forums is hard!

When the forum was down the website was OK from my location.

For some reason only the forum was unreachable. I've tried a traceroute forums.freebsd.org out of curiosity, and after a while the packets were keep jumping between the same 10-20 IPv4 addresses, and nevel live that class of IPs, but I didn't saved the results. Made me think the cause was bad routing.

Also, "perina" in Romanian is a popular form for the word "perna", both meaning "pillow". Romanian is mostly a Latin language, similar with Italian, French or Spanish, so it is rather rare to find common words between the Russian and the Romanian tongues.

I didn't know about the Russian "perina" before, seems like FreeBSD will teach me way more than I ever dreamed of! :)
 
Yeah, let's just pretend that nothing happend.
OK I admit that making a PR was appropriate.
All the silly chatter below it was more than "Reporting a problem".
State your problem. Then state your solution. If you have no solution to offer than stand down.

I also think its silly to expect a public response.
All important decisions would be made in private between the relevant parties.
We don't manage by FreeBSD by survey.
Not in total secret but allow the negotiations to be carried out in private.
Maybe we sweep this under the rug because it was a rare occurrence.
BFD
 
In this case, it wasn't Chaka Khan. It wasn't Ozzy, either (and the vaxorcist has retiread anyway). Long story short, the dark chamber had to call for support from the laundry and the 13.th directorate...
Delighted to see that someone remembers the VAXorcist.
 
I wonder why that PR is still "In Progress" …

My bad. The bug report was, essentially, opened for the outage, for tracking purposes. It might have been useful for more, during the outage, but as things turned out: there was no need.

The report could/should have been closed by me immediately after normal service resumed. It's now closed, fixed.


Hey eternal_noob good morning (or whatever the time is in your zone – I'm unsure about what's mapped), let's have a virtual cup of tea and put our feet up for a moment. It's quite unlike you to exaggerate. You're amongst the people who are, naturally, respected for level-headedness and balance. Take a sip of that cup of tea, take a sip of the respect.

let's just pretend that nothing happened. …

There's no such pretence. Something, or things, happened.

… people don't want to discuss but shut down communication. …

Honestly, I see no attempt to shut down.

Readers, please, don't misinterpret my closure of the bug report:
  • downtime ended – fixed
  • if there's more to be done, it will be out of scope of that particular bug.



Here, we're eight pages into discussion, which is not necessarily a bad thing, however:
  • a slight increase in length can greatly increase the likelihood of people overlooking, or ignoring, relevant information
– and I'm amongst the people who overlooked at least one key point.

Healthy scepticism is good. Repeating the scepticism is, IMHO, unlikely to make the scepticism any more widespread.

eternal_noob please know that you're not alone in taking a cautious approach to things. For some such things, I'm unapologetically private, and no amount of demanding openness will create openness. We're all human. Push a human too hard or too fast, there'll be the opposite of what's wanted from the push.

… allow the negotiations to be carried out in private. …

+1 to that, and more; and this upvote is not intended to end discussion.

… this isn't the kind of shop that makes "statements" every time something breaks. There was a machine in either the ports or freebsd-update round-robin DNS that was out for months …

I don't recall that incident but certainly, there are package infrastructure bugs.

It's possible that one of the package infrastructure bug reports coincided with a report of a bug in or around another FreeBSD infrastructure. With all due respect:
  • to identify any such coincidence, and then begin worrying whether each coincidence is symptomatic of a breach (or whatever), would be a waste of time.
 
I don't recall that incident but certainly, there are package infrastructure bugs.
It was like ftp5.something.freebsd.org or something. It would get stuck in the DNS cache and I'd have to go into some config file to hardcode the tool to use ftp1 or something because it would just time out.
 
It wasn't a big deal to me. It was a few days, I noticed it, but thought it was on my end or part of the Internet infrastructure.

However, I realized that for someone needing to solve a problem to install FreeBSD or fix a problem, it would be a big deal. Mostly new users, prospective users or those doing something on a high learning curve. That's Internet traffic lost, and people who get discouraged from using FreeBSD. It's likely a low percentage of users who really need the forums, but every little bit is something. For most on here, who've solved the issues they tried to fix, it's not a big deal.
 
… is this datacenter for FreeBSD forums only?

It might be not as grand as a data centre.

… It started off on someone's personal host and it's been running there ever since. …

Wherever it is, physically or virtually, it is (according to this description):
  • personal
– and many people benefit from the person's arrangement. Again: I thank this anonymous donor.

I am a person with ASD, so hello ;)

Is that autism spectrum disorder? Tell us more, here:


… problems with 'pkg upgrade' also, …

… more info. Probably pkg have …

I don't imagine any connection between:
  • the personal host for FreeBSD Forums
  • infrastructures for packaging.
A separate topic will be good, thanks.


Welcome to the forums! 🎉🥳

+1
 
It was like ftp5.something.freebsd.org or something. It would get stuck in the DNS cache and I'd have to go into some config file to hardcode the tool to use ftp1 or something because it would just time out.
A quick Google search on the issue turned up this gem from the FreeBSD dev mailing lists:
FreeBSD doesn't keep a local cache of DNS lookups, so there's nothing to do.
This info is a little old, though - from 2016.
 
When using RR DNS, my understanding is that the response will get cached somewhere for a time, often your router. Either way, I would run dig on ftp or whatever is was and whenever it broke, it would reverse to the same one: ftp5, or something like that. I don't remember exactly.
 
… Internet traffic lost, and people who get discouraged from using FreeBSD. …

Possibly. There's also this:

1642843519689.png


I suspect that we're enjoying a small invasion of new members as a result of curiosity during the downtime.
 
As the name of the thread may suggest, "Life without FreeBSD Forums is hard!" it's more or less about discussion (read complaining ;) ) about it and its consequences, not only about solving the issue, for relevant only information we should probably start another thread ;) The fact that moderated posts are posted according to the time of sending, not the time of acceptance (which changes the order of posts, and makes easy to miss something out) doesn't help either.

In Unix world, silence means "everything is OK", and everyone (well, at least majority) agrees that it wasn't OK. In such case we could expect at least an error code, and there's none, and it causes some disonance, because it *may* suggest that people responsible for the FreeBSD don't share the same UNIX mindset as its users, which is worrisome since they (developers) are responsible for its direction. I think we have a right to ask questions, and search for the answers. At the moment we know *nothing*, so it's hard to say which info is relevant, and which is not. FreeBSD forums are *official*, so they should be treated as part of the official infrastructure, no matter where they're hosted.

Trying to understand is never a waste of time. I'm not saying it was a breach, I'm not saying it wasn't a breach. As for my problems the most likely pkg wasn't tested with such breakage in mind, and something was overlooked and pkg failed. But how can one (help to) solve the problem without understanding it first? And first step is to rule out the obvious - that the both problems were connected.

Also, "perina" in Romanian is a popular form for the word "perna", both meaning "pillow". Romanian is mostly a Latin language, similar with Italian, French or Spanish, so it is rather rare to find common words between the Russian and the Romanian tongues.

I didn't know about the Russian "perina" before, seems like FreeBSD will teach me way more than I ever dreamed of! :)

There's also Perun (Pierun, Piorun, Pieron) of Proto-Indo European origin, god of thunder and lightning ;) *per- "striking" is used in many polish words, like pierze - feathers, often used in pillows; pierzyna - bed sheets; prać - to wash clothes, in the old days people washed clothes by striking them with a stick; parzyć się - to make love; parzyć - to burn; naparzać się - to strike each other. The most popular and widespread* polish word is probably wampir (and similar upiór) - vampire, originally it was probably someone strucked by a lightning.

*well, I can think of more popular word, but it's unsuitable for this forum ;)
 
Sometimes, it helps to connect some dots and dig around a bit. I think this thread does provide a good example of how to do that. Relating outages to weather around the datacenter and the other nodes on the Internet, and realizing that the weather can and does get in the way of communicating (Even asking a pro to get up and fix it can be a challenge if the rain/snow created mudslides that carried the cell tower away) - that's just a couple of the dots we gotta be able to connect. One would think that we're invincible, and have this global reach, thanks to the Internet - but in reality, the Internet is a pretty fragile source of information, very susceptible to weather. Yeah, we have the brains to maintain that network, build in some redundancy to ameliorate the impact of weather, but no, the Internet is not the weather-proof, invincible behemoth that we think it's supposed to be. 😩
 
Here's what I think happened.

Some time back, John said, "I've got space on my host system and you can use it for the FreeBSD forums." Great! "But don't ask me to do anything with it. This is up to you." Ok. Great! Thanks!

The system went down. A few hours later, someone emailed John about it. John was away or doesn't really look at his email that much. (He might even be retired.) It's not that important to John and he doesn't notice.

Whoever emailed John figured John would take care of it and went about their own business. Nothing happened. A day goes by and someone notices. Emails or texts John again. John's busy or doesn't notice the email/text again.

Three days go by and John finally notices. He reboots the system and all is well. John goes back to what he was doing.

And that's the story (I think).
 
Because people don't want to discuss but shut down communication. Sad.
I don't want to shut down anything. I am just a regular joe with an opinion. Just like you.
But when you used the word 'ignorance' in bullet point 3 I had to get loud rude.
Mister @danger has feelings as does everyone who works this forum for FREE.
Insults are not appreciated for our volunteers.
3. It took 3 days to recover the forums (WTF? Big problem or just ignorance?)
 
I respectfully disagree; there is information, if we pay due attention to the authors.

Yes, there's some information so we can create some hypotheses, but we know nothing for sure. Just a moment ago in Alerts I had a bunch of "later_today_at_x" instead of time, so obviously something still isn't working properly.

But pkg didn't fail. It refused to install a corrupt package as it should have. The network failed.

That would be true if there was a breach or the entire network failed. It didn't, and from what I know, pkg should automatically choose a working mirror, so it failed at this task. It also had problems with proper numerating of downloaded packages, this is (should be) independent of the network connection.

And to astyle: Yes, of course it's possible, but just because something is possible doesn't mean it really happened. What's more, even if probability of something equals zero, it doesn't mean it can't happen.
 
Back
Top