Hello Mastodon 👋 I’m a Redis + NoSQL enthusiast who loves building fast, scalable data systems.
When I’m not optimizing latency, you’ll find me editing audio/video projects or out on the motorcycle chasing curves. #Redis#NoSQL#Motorcycles#AudioVideo#Introduction
Hi, in Mastodon v4.4.8+glitch I've many sidekiq errors in tab "Retries":
Redis::CommandError: NOPERM No permissions to access a channel
It happens with and without Redis' password options (and in 2 instances I've set up, v4.4.8+glitch and v4.5.0-beta.2+glitch).
Had in first instance a problem with Sidekiq's default queue before (many queue'd entries), buit I could solve it with a second default queue process yesterday.
Any idea how I can stop this? There are 290k retries.
Hey folks, if you run Redis you should be aware of a CVSS 10 vuln, CVE-2025-49844, which is a lua related RCE. Redis have release a patch for this and 3 other CVEs. According to Wiz, this vuln has existed for 13 years. That means forks such as Valkey may also be impacted. Valkey has also released updates to address the same CVEs.
I missed this news a couple of days ago: #Redis is officially FOSS again, as it has backtracked from #SSPL and moved to AGPL (which IMHO is what they should have adopted all along).
This closes the cycle of my prediction:
That Redis had made a terrible mistake by switching to SSPL
That the switch to SSPL would have just resulted in a fork fever that would have critically eroded their user base and reputation
That those forks would have been eventually included in standard Linux distros instead of Redis itself because of their OSI-compatible licenses, and adopted as alternatives to Redis by the cloud providers (hence invalidating the whole point of SSPL)
That Redis would have eventually backtracked and switched back to AGPL
There's a 5th point in my prediction that hasn't materialized yet though: that the damage is already done and it will be hard to revert.
Fedora, Arch and Alpine have all replaced Redis with Valkey in their repos already, and Debian has opted to temporarily provide both.
Many Docker configurations, cloud and on-prem deployments, Terraform templates, home servers etc. have already all migrated to Valkey.
Not only, but the seismic fork allowed Valkey to implement features that have been requested in Redis for a while but not implemented because of the most complex governance of the project (like RDMA). Plus, while Valkey guarantees parity of features as of Redis 7.2.4, the mismatch of features is only supposed to increase as the hard fork diverges. This makes it even more unlikely that an enterprise software that has already migrated to Valkey takes the risk of migrating back to Redis later.
The other project that still partly uses SSPL is Elastic. And even in that case the decision has backfired. It opened the way to AWS to release OpenSearch under an Apache license, and grab a lot of FOSS developers to power its proprietary cloud efforts for free instead.
#MinIO is the new #Redis, software updates and launches, a spotlight on #LogForge -- a #log monitoring platform, and more in this week's #selfhosted recap!