CloudLinux OS repository mirrors: updates fail with slow or unresponsive mirrors

Incident Report for CloudLinux

Identified

We are sorry that this update comes later than the 16:00 UTC we promised

Mirror list: in our checks from 18:15 to 18:21 UTC the default mirror list returned seven mirrors in Europe, the US and Asia. Six of them served CloudLinux OS 7 repository metadata at 1 MB/s or faster, and one was still slow. From 18:10 to 18:13 UTC four or five of the six returned a 404 error for this metadata

Second issue: since 16:06 UTC we have received reports of a separate problem. yum and dnf intermittently fail to download the metadata of repositories served from repo.cloudlinux.com, for example cl-ea4 and cloudlinux-updates-testing, with "Status code: 404" or "Curl error (18): Transferred a partial file" errors. The error can name the host repo.cloudlinux.bunny.cl-mirror.net, to which repo.cloudlinux.com redirects. Our infrastructure team has identified the problem and is working to reduce the impact. The fix has not been applied yet

If an update fails with a timeout, run "yum clean all" (CloudLinux OS 7) or "dnf clean all" (CloudLinux OS 8 and 9) and run the update again. If it fails with one of the two errors above, clearing the cache may not help. In our checks the same download failed and then succeeded a few minutes later, so please run the update again later. If it keeps failing, please open a support ticket and attach the yum or dnf output

We will post the next update when the fix for the second issue is applied or when our infrastructure team shares new information
Posted Nov 19, 2026 - 18:31 UTC

Monitoring

The mirror list has been fixed and again lists several mirrors in Europe, the US and Asia. Our checks at 14:23 UTC downloaded repository metadata at normal speed from the mirrors in Europe, the US and Asia. If a server still fails or downloads slowly, run "yum clean all" (CloudLinux OS 7) or "dnf clean all" (CloudLinux OS 8 and 9) and run the update again. We are monitoring the mirrors and will post the next update by 16:00 UTC.
Posted Nov 19, 2026 - 14:30 UTC

Investigating

Since 12:12 UTC on 19 November we have received reports that yum and dnf on CloudLinux OS 7, 8 and 9 servers fail to download repository metadata or download it very slowly. The default mirror list (cl-mirrors), including the ELS list, sends servers to mirrors that are slow or not responding, so updates fail with timeouts or "all mirrors were tried" errors.

We have escalated the issue to our infrastructure team. The problem is in our mirror list, so no changes are needed on your servers. We will post the next update by 15:00 UTC.
Posted Nov 19, 2026 - 14:17 UTC
This incident affects: CloudLinux Network.