Donwside to 2: Your VM becomes harder to move between hardware, you lose snapshotting capabilities from a copy-on-write image.

5 is flexible, but has limitations. For example you wouldn’t want to run databases on NFS volumes.

If initialization time is the only problem with 4, you could create several smaller images on the disk. Create the first one, initialize the VM and set up an LVM volume on it, then start creating more volumes and extend the LVM volume.

What REALLY irks me is that the gboard keyboard on Android is context-sensitive. No, I don’t want to have a shortcut for “.com” when long-pressing the period key while typing an address, I want to type a fscking dash!

Betteridge’s law of news headlines strikes again!

As in contract law, the solution is to eliminate overarching or vague promises. Instead of promising to “balance the budget”, have them produce a budget plan. Instead of promising elections reform, promise election reform pilot programs. And let’s not kid ourselves, election promises made in good faith are a rarity these days. It’s time to make it harder to lie to the electorate.

Punishing the breaking of election promises would be a start. Those are not ambiguous or unintentional, and it should be punishable as a breach of contract.

The alternative would be a non-standard diaper app that, rather than hiding the incoming call, would pick it up and drop it. I don’t know if such software exists.

I assume you meant dialer app 😆 . But anyway, for some Android phones you can use call screening.

Good article for discussion.

Health checks is one situation where kubernetes really shines. It makes a clear distinction between readiness probes (when the pod is ready to start serving traffic), liveness probes (when the pod should be considered dead), and startup probes (when the pod has finished bootstrapping). Coupled with autoscaling it then becomes acceptable to have a pod stop serving new traffic when it’s too busy, because other pods can be created in a short time to take the extra load.

Including backend checks in your application depends on its nature. I think the mistake that the article’s author made was not to include the checks, but to have too big of a blast radius when the check fails.

Argentina lost their first game last year, and they went on to get the cup. Canada still has a chance.

I think the foreach one should have been recursion.

Both are concerning, but as a former academic to me neither of them are as insidious as the harm that LLMs are already doing to training data. A lot of corpora depend on collecting public online data to construct data sets for research, and the assumption is that it’s largely human-generated. This balance is about to shift, and it’s going to cause significant damage to future research. Even if everyone agreed to make a change right now, the well is already poisoned. We’re talking the equivalent of the burning of Alexandria for linguistics research.

The RV260 supports SNMP. You can use that with a network monitoring tool of your choice to get ifInOctets/ifOutOctets data. The rate of change on those numbers is then the amount of traffic sent/received.

$50M is pocket change to them. Needs more zeroes at the end.

Spez: this will blow over Also spez: this cannot be allowed to continue

Link goes to the wrong page. I think you meant to post https://globalnews.ca/news/9767962/ontario-first-tornado-2023-confirmed-talbotville/