LWN Downtime Due to Linode, Not LLM Bots
Yesterday: LWN Has Been Down for a Long Time, Another Casualty of LLM Bots?
So LWN is back online, with Mr. Corbet explaining: "The good folks at Linode still have not managed to fix whatever broke in their data center, so we are running on an emergency backup server. Things seem to be working, but the occasional glitch is to be expected. Please accept our apologies for the extended downtime!"
He later added: "we're back on the regular production server, and all seems stable now."
Mike Rockwell spoke of potential for data loss: "I’ve been a happy Linode customer for years, but they experienced an outage Sunday morning that took my Cloudron server offline, impacting my Mastodon and Pixelfed instances. As of this writing they’re still offline and I’ve received an email letting me know that there is a potential for data loss. Luckily, I have backups of my entire Cloudron server in case I need to restore. I’m going to wait a bit longer, though, to see if the team at Linode are able to get everything up and running again."
One commenter uttered: "I’d be _extremely_ curious to read Linode’s postmortem for this outage."
I read all the other LWN comments, some of them make suggestions for a more robust setup. The last one says: "I agree with this. And it's another reason why I prefer to get just generic Linux machines or VMs rather than paying for fancy cloud stuff. There's much less vendor lock-in if all you need is root on your own Linux machine; it's cheaper; and you can manage your own backups. Don't fall for the siren song of fancy cloud APIs that tie you to a specific vendor. They're convenient, but they hook you in."
Clown computing is a lot of hype, much like slop, and the costs are typically hidden.
The above speaks of "generic Linux machines or VMs", which is what we have for this site. We're OK as long as it can boot OK on any server. █