Why is the site so often unreachable?


Today, some minutes ago, the network connection was cut and the site unreachable.
Since the switch to this forum it is the third (or fourth?) time, the site is down or could not connect.

Is this a generel problem of the hoster or software or what?

Would this be better if “we” switch to another hoster?
Or increase the bandwidth?

What else may help?

Thanks in advance


The same here…for a few minutes though.


A few minutes of down time once a month or so is hardly worth getting worked up about.
Want to improve things​? Maybe reach for your​ PayPal account and pay for a year of hosting.


Yeah you’re right. I feel embarrased now for complaining this…


Cos I simply update our servers time to time and therefore I’ve to shut them down.


Thanks for the clearifying - i thought not that this could be so simple :slight_smile:

How about an extra section “Website” in the “Announcement” category and a fast post when maintainance and/or downtimes are upcoming/needed?

This will avoid future questions and give the community the direct thread to react if a problem comes up after a update of the forum software or website.

Thanks for your work Philm!


If you’re still connected to the forum you’ll get a maintenance or error page. If you try to connect to it during maintenance you get the standard not reachable error. This is how Discourse is designed. Sure I may go thru the hassle and redirect the URL, but for a 5 minute downtime this would be to much effort. Also I don’t plan those downtimes. When I get an email notification and the time is ok, I’ll simply do it. Sure we can improve that process …


I’ve complained about this in the past (more than once): The dev’s do an awesome job of being dev’s, they just don’t do an awesome job of communicating with their users.

Phil communicates better than most dev’s do, even so, we don’t get told important things sometimes, until after they have happened.

Other important things are never really explained clearly.

Anyway, that’s just my bitch re. Manjaro (I don’t think that I really have any others… - oh, apart from some of the problems with Discourse that is). :wink:


I don’t think the devs need to answer to the public, it’s there job to maintain things not work around user’s schedules.


It would be nice if there was someone who is a part of “the team” on the other side of the curtain, who had good communication skills, & had a desire to keep the forum community up to date on the goings on. When an important change is coming down the line & when one does, they could inform us in good easily understood english (for this forum) that new users would be able to understand.

philm does a good job re. updating us with what’s going on with the packages being updated/upgraded, added, removed, modified & such.

From time to time other stuff happens that doesn’t get communicated as well as many of us would like (or at all).

As you say PR, that’s our problem. Though that doesn’t remove the fact that there really is plenty of room for improvement on the communications front as has been shown in the past (in spades!).

This has been a problem with Manjaro since I joined the community in late 2011.


Today July 4, was another day of looooong connection days, and frequent time outs.

I just chocked it up to too many North Americans on holiday and flooding the forum. High Server Workload.


I am now in Indonesia using a 4G connection for both the smartphone and the laptop (tethering). Only the Manjaro pages take a hell of a time to load. Other internet pages from around the world come in at normal speed. This low speed is only today, until yesterday it was still okay. What could be the cause of that?


I’ve also had that problem lately too. It is an inconsistent problem, sometimes I just cancel waiting & go to another site.


It looks better at this moment. Yes, I also went to other sites when the connection to Manjaro did not work.
Hope it will stay better now.


It’s hard to deny the fact that these discourse/hosting i/o errors have been going on for over a month now, give or take a few days. It doesn’t matter if it’s a themeing-plugin to blame or whatever, -The point is, it’s annoying at best, at worst, it will turn people away from here, eventually.
Discourse needs to FIX this.
That said, I’m almost leaning towards a FluxBB solution, instead of these continual Discourse hosting i/o errors.
The squeaky wheel gets the grease, so they say, but until then we’ll ALL have to put up with this, no matter how much Discourse wants to bury these issues… <- IMHO.

No, I’ve never seen these many 504’s as I have lately, so no, it wasn’t that bad before, atleast as I remember, especially that far back, well, maybe early on yes, but not like this.


Tuesday and Wednesday were especially bad for me but even normally it is a bit slow.


That’s not what Handy was referring to - it was to a lack of communication surrounding changes and maintenance.

As for the rest - this sort of unreliability is really pretty poor. :frowning_face:


I understand exactly what Handy said, although I did specifically point to what i quoted. And Handy can answer for hinself.
Anyway, I’m sorry we’re all having these Discourse issues too, but sooner or later, somethins’ gotta be done about it.


Is there any proof this is even a discourse issue?
All the messages I see are nginx messages, and they look a lot like out of resources than anything else.


It’s most likely a server resource issue. I suspect the server is hitting its RAM limit and swapping to disk, hence the timeout errors.

The other issue is that the Discourse instance is hosted on Digital Ocean; they generously provided a substantial initial credit which would normally last a long time. I think Discourse is heavier than was originally anticipated so the droplet has already been upgraded once; upgrading it again will burn through hosting budget quickly. My suspicion is this is why it hasn’t already been done (and once you upgrade the droplet I don’t think you can downgrade it again… though there should really be some bursting available, that’s the whole point of “cloud” servers).