Skip to main content

Redis 8.10.3

Release Date: October 26, 2026

Redis 8.10.3 landed on October 26, 2026. Let's not get carried away with a victory lap. The entire changelog is one bullet, and it is filed under Update urgency: SECURITY.

The note reads plainly enough: a heap buffer overflow when propagating a PUBLISH or SPUBLISH message larger than 4GB over the cluster bus in cluster mode.

That is the release. No new commands, no config additions, no throughput work. Just a memory-safety fix in the cluster bus, shipped the same day across four branches: 8.10.3, 8.8.4, 8.6.8, and 8.4.8 are all dated October 26 with identical wording.

Why You Still Upgrade

It is tempting to read "larger than 4GB" and conclude the bug cannot reach you. Fair instinct, wrong conclusion. A single message that big is unusual, but the fix is in code that handles message propagation between cluster nodes, and the branch pattern tells you the maintainers considered it serious enough to backport to 8.4. Memory-safety bugs do not politely respect the assumptions you made about your data sizes.

Cluster operators should treat this as a mandatory patch. Standalone and single-node deployments are outside the described path, but there is no version of this where running 8.10.2 for another month is the smart call.

The Caveats

First, the notes name no CVE. If your security review process is keyed to CVE identifiers, you will need to open a ticket manually or wait for the advisory to be published.

Second, this is a server restart. There is no hot patch and no CONFIG SET that avoids it. For a cluster, plan the rolling restart properly: one node at a time, wait for the cluster to report healthy, then move on. A 4GB buffer overflow fix is not the moment to discover that your failover takes longer than you assumed.

Third, 8.10.3 is a patch on top of the 8.10 line, and the previous release, 8.10.2, only arrived on September 19. Two releases in five weeks on the same branch is not unusual for this project, but it does mean your changelog review should cover both if you skipped the earlier one.

What to Actually Do

  • Inventory every Redis deployment and record which branch it tracks.
  • Prefer the newest branch your client libraries and modules support. 8.10.3 if you are on 8.10, 8.8.4 or 8.6.8 or 8.4.8 otherwise.
  • Drain and restart during a window you control, not during an incident.
  • Verify after restart that the cluster state is stable and replication lag is back to normal.

No drama required. One line of changelog, one restart, and a bug class that stops existing on your nodes. Do it this week rather than next month.

Other Versions

We use cookies for ads and analytics.