7

Moderator's Note: This is an active, ongoing issue. PLEASE do not post answers on this question saying "me too" or such; they will be deleted.

Because this is an ongoing issue, this question has been protected to prevent brand new and low-rep users from posting answers on it.

This is the canonical post for "I'm having issues with updates" today—if the issues are able to be determined as directly related to the mirror issues today.

Like others (1, 2, 3, 4) we're experiencing problem connecting to http://archive.ubuntu.com

I tried using servers from different providers in UK, France, Netherlands, Ireland and US; same story: HTTPS works, but under HTTP some IPs respond very very slowly, while others don't respond at all.

dig +short archive.ubuntu.com | sort -V
91.189.91.81
91.189.91.83
185.125.190.81
185.125.190.82
185.125.190.83
for ip in 91.189.91.81 91.189.91.83 185.125.190.81 185.125.190.82 185.125.190.83; do \
    printf "========> ${ip}: "; \
    start=$(printf "%(%s)T" -1); \
    if ! curl --connect-timeout 60 \http://${ip} -H 'Host: archive.ubuntu.com' -o /dev/null -s; then \
        printf "Failed "; \
    else \
        printf "OK "; \
    fi ; \
    now=$(printf "%(%s)T" -1); \
    echo "$((now - start)) secs"; \
done

UK

========> 91.189.91.81: Failed 60 secs
========> 91.189.91.83: Failed 60 secs
========> 185.125.190.83: Failed 60 secs
========> 185.125.190.82: OK 35 secs
========> 185.125.190.81: Failed 60 secs

France

========> 91.189.91.81: Failed 217 secs
========> 91.189.91.83: Failed 60 secs
========> 185.125.190.81: OK 14 secs
========> 185.125.190.82: Failed 227 secs
========> 185.125.190.83: Failed 60 secs

Netherlands

========> 185.125.190.83: Failed 221 secs
========> 185.125.190.82: Failed 60 secs
========> 91.189.91.82: OK 0 secs
========> 91.189.91.83: Failed 157 secs
========> 185.125.190.81: Failed 221 secs
========> 91.189.91.81: Failed 184 secs

Ireland

========> 91.189.91.81: Failed 162 secs
========> 91.189.91.83: OK 88 secs
========> 185.125.190.81: Failed 116 secs
========> 185.125.190.82: Failed 119 secs
========> 185.125.190.83: Failed 60 secs

US

========> 91.189.91.81: Failed 227 secs
========> 91.189.91.83: Failed 80 secs
========> 185.125.190.81: OK 44 secs
========> 185.125.190.82: OK 53 secs
========> 185.125.190.83: Failed 104 secs

I don't have contact with Canonical; is there anyone here that can let them know?

jtheo
  • 71
  • 4

3 Answers3

6

Ubuntu service outages can be tracked on https://status.canonical.com/ . At least from my perspective the problem contacting Ubuntu's default apt servers appears to be over (2025-05-29T18:44+00:00) but https://www.reddit.com/r/Ubuntu/comments/1kxyig1/anyone_know_what_is_going_on_with_ubuntu_archives/ seems to suggest the issue was present at 2025-05-29T01:32:32.129Z so perhaps the problem comes and goes...

The issue would manifest in a variety of ways:

  • Extremely slow downloading of update lists and the packages themselves
  • Requests failing with 403
  • Outright failures to connect to the mirror in the first place

This could be especially problematic when downloading lots of packages in a stateless way (e.g. when building docker containers).

While the issue was occurring workarounds suggested by the Internet seem to be:

  • Try using a mirror (e.g. by changing both archive.ubuntu.com and security.ubuntu.com to another host within /etc/apt/sources.list): see https://launchpad.net/ubuntu/+archivemirrors for a list or try using the mirrors on AWS (e.g. us-west-2.ec2.archive.ubuntu.com) or Azure (azure.archive.ubuntu.com)
  • Try changing http to https within /etc/apt/sources.list for archive.ubuntu.com and security.ubuntu.com
Anon
  • 382
4

I have contacted my contacts at Canonical through other means, and they checked with Canonical IS and were able to confirm at least this so far:

Canonical IS is aware of the infrastructure issues related to the mirrors, and are working on it. However, Canonical does not yet have any ETA on a fix for this. You are encouraged to test other mirrors that might work, or simply wait for this to be fixed.

(Last updated: 2025-05-29 13:49 UTC-04:00)

Thomas Ward
  • 78,878
1

Personally, I've set my deployment to use a mirror provided by a cloud ISP in my country (since the great firewall of China often slows down the connections to IP addresses that aren't censored). So personally I'd suggest try cloud ISPs in your location, but of course it's likely their content is outdated by this point.

Interestingly, when I issue update commands on my deployment, I see it download upgrades from the domains of my local mirrors automatically, and not from the archives' domains. Not using the Kylin edition though.

Update 2025-06-04, the issue seem to be fixed now.

DannyNiu
  • 127