For me, nextcloud.com is down, the site is not loading at all
1 Like
@jospoortvliet mentioned above, that they are looking into a CDN solution but it will take some time. At the moment we have to live with the current situation and make the best out of it, even if that is not the answer people are looking for 
and make the best out of it
I already do, see: *.Nextcloud.com - connectivity issues - #52 by mritzmann
Okay, here’s an out of scope idea. With an installed base of millions, show a one time popup on the next update asking for help (this would be exposed in the settings as well for later opt-in). The request would be to set aside, say, 500MB of space for an Updates folder to be shared with others, and the owner could set a maximum number of downloads (e.g., 10, 100, unlimited). Fresh updates are fetched via cronjob that checks 3 volunteer servers before defaulting to the root server, in case one of them has already fetched it. The hash of the downloaded file is checked from the root before the upgrade is made available to others.
When the automatic update function is called, it randomly tries 3 volunteer servers in the current list before falling back to the root server.
Instant volunteer CDN.
Even if you put it before to update to your dowload folder?
Downloading: downloads the code in the version it should update to. This is also shown in the web UI before the update is started. This archive is downloaded to
/updater-INSTANCEID/downloads/.
I strongly believe if you download archive manually e.g. via torrents and put it into downloads folder, the automatic updater will simply skip this step and proceed to “Extracting” step. Maybe it will calculate some checksum before
No 
(updater/lib/Updater.php at a4a3248d541fb82f3c9f8f9bf0774c1b70324ab3 · nextcloud/updater · GitHub)
But sounds like a good enhancement.
1 Like
There are existing repos for open source software (https://www.gnu.org/server/mirror.html), either use one of them or set up something similar. Setting up a similar mirror with rsync shouldn’t be a that difficult either.
Yeah, I also checked it after post and was laughing on my broken expectations: ->recursiveDelete 
could be verifyIntegrity instead.
But sounds like a good enhancement
I still think this should be the besser enhancement: Possibility to use a local Mirror · Issue #291 · nextcloud/updater · GitHub
The result is the same, but in beautiful 
@mritzmann
You can also download the newest version of nextcloud from another source and then copy it to “nc.zip”
Read /setup-nextcloud.php says it has succeeded but it hasn't installed anything - #2 by devnull
The problem: there is no mirror
How it works with updater? Seems its removing file if presented.
I don’t think so. I care about updates with updater.phar not about new installations. And there’ s no nc.zip in the code. This is the reason why I created this issue.
Tonight the apps.nextcloud.com site literally was unreachable (independent of DNS provider) nextcloud.log gave me the clou.
To make sure, you can try to reach the url from server as well as typing the url directly in your browser. If DNS is broken only one should fail and the other one succeed. If it is the site, both will fail.
(edited)
I noticed the same behavior on multiple nextcloud instances.
In our forums (nethserver) we also had members that noticed the issue. Glad it has been solved.
Nextcloud archives from download.nextcloud.com have been very slow (~100kb/s) for at least the past 20 hours.
Is this issue known already or where should I report it?
I observed the same from two different networks, therefore I suppose that this is due to the source network/host.
1 Like
Nextcloud 20 has been officially been published at the weekend so that many users are currently downloading the latest version. Therefore it is very likely that the servers are a little bit overloaded at the moment 
CC: @jospoortvliet
2 Likes
That is a reasonable explanation, indeed. I was aware of the update but expected Nextcloud to have better bandwidth, therefore the only explanation I thought of was a technical problem.
1 Like
This leads to Errors while updating though. I am permanently stuck at Step 4. Each time I restart the update (with the button that appears if an error e.g. due to NGINX timeout), the perfectly fine file sitting in updater…/download is discarded (although I verified that it’s complete and fine using sha2556sum). Therefore, the updater app should be enhanced – this would save bandwidth on the server for all users that do retry!
Every release the same problem.
I can confirm this behavior: The download indeed finishes successfully, but takes way longer than the updater is willing to wait (times out). The downloaded file is still perfectly fine, matching checksums and all. So, being stuck at Step 4 (or occassionally even Step 5, when apparently the checksum file is not reachable for download), the perfectly fine update file gets deleted and downloaded anew.
So, the overloaded download servers are the update script’s own naïve fault. Something like this should be avoidable so easily.