Jim Salter

Things Jim Says on Podcasts

Covers home and small-business IT infrastructure, from ZFS storage and network gear to backup practices, usually judging products by how they fail. That shows up in his comparisons of Wi-Fi controller brands, where he weighs how badly cheap hardware degrades once the management layer goes down. Hosts the podcast 2.5 Admins and writes for outlets including Ars Technica, where his ZFS and storage coverage is widely read among sysadmins.

Where to Find Them

Jim Salter hosts 2.5 Admins . They have also been a guest on Late Night Linux (5 times) and LINUX Unplugged (3 times) .

Recently: “2.5 Admins 326: *sshole” on 2.5 Admins (November 2026); “2.5 Admins 325: Fmail” on 2.5 Admins (November 2026); “2.5 Admins 324: Less Canadian” on 2.5 Admins (November 2026); “2.5 Admins 323: No /boot For You” on 2.5 Admins (October 2026); “2.5 Admins 322: Internet Defrigeration” on 2.5 Admins (October 2026); “2.5 Admins 321: Backup Strategy” on 2.5 Admins (October 2026).

What They Said

“Flock described this as a one in a million error, which is hilarious given that the exact same thing happened with the same entry in the NCIC database halfway across the continent one month later. But even if it was a one in a million error, Flock readers scan 20 billion times per month. So a one in a million error means 20,000 of these errors a month.” — Jim Salter, 2.5 Admins

The hosts were discussing a driver surrounded by police after a license-plate reader matched a dealer plate entered wrongly in a national database. Jim Salter answers the vendor's claim that such errors are vanishingly rare.

2.5 Admins · 2026-11-19 Permalink → Listen →
2.5 Admins Around 20:30 into the episode
Allan Jude

That's a great batting average they have there. Yeah,

Joe Ressington

that's not quite the one in a million that Flock claims.

Jim Salter

That's true. It's not the one in a million that Flock claims, but let's talk about that one in a million. Flock described this as a one in a million error, which is hilarious given that the exact same thing happened with the same entry in the NCIC database halfway across the continent one month later. But even if it was a one in a million error, flock readers scan 20 billion times per month. So a one in a million error means 20,000 of these errors a month. So we said that we weren't going to talk about politics, and we're not. What we're really talking about here is data handling. And I think anybody who has functioned as any kind of an information analyst, a sysadmin, a DBA, should already be horrified at this very imperfect storm of every possible error glommed together. You've got a database that is designed to be as sticky as possible with all…

Joe Ressington

Let's do some free consulting then. But first, just a quick thank you to everyone who supports us with PayPal and Patreon. We really do appreciate that. If you want to join those people, you can go to 2.5admins.com/slash support. And remember that for various amounts on Patreon, you can get an…

Jim Salter

Only one. If you're working with the team and you want to be able to easily find team members with direct experience with your virtualization platform of choice, you're going to find a lot more folks who already are familiar with using Proxmox than are already familiar with running libvert…

Allan Jude

Yeah, I think to your point, Proxmox is great if you need clickOps, right? If you need you or other people to be able to do it with just a browser and point it at it and so on. But I think especially for libvert, if it's, I want to use this machine I already have, or this machine has other jobs and…

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“If you're not testing your backups regularly, you don't know if you have backups or not. You've got a bunch of crap. It might be a backup. It might not. You won't know until you need to use it.” — Jim Salter, 2.5 Admins

His co-host had just said the key lesson was to always test your backups. Jim agrees, and argues that regularly testing ZFS replicas reduces the need for an expensive extra backup chain.

2.5 Admins · 2026-11-05 Permalink → Listen →
2.5 Admins Around 28:31 into the episode
Jim Salter

…it work? And you go, yep, this all looks good. I've tested it in the ways that I feel like I need to to assure myself this database is not corrupt and I can blow away that VM and I know that was good. And next month or next quarter or whatever your schedule is, you can test it again to make sure.

Speaker 4

That's probably the key lesson here. Always test your backups, no matter where you go.

Jim Salter

Yeah, I mean, if you're not testing your backups regularly, you don't know if you have backups or not. You've got a bunch of crap. It might be a backup. It might not. You won't know until you need to use it. And if you are testing your backups, the importance of that super expensive, hard to pull off, you know, extra backup that relied on your one ZFS replication at work in the first place, it's not so big anymore, you know, especially when, again, you consider that those ZFS backups, it's not just a backup. Backup each one of those additional chains, it's got archive depth. You know, it's not just…

Joe Ressington

Right. Well, we better get out of here then. Thank you very much for joining us, Sean. It's been great having you.

Speaker 3

Thank you for having me. I'll see you guys next time.

Joe Ressington

Remember, show at 2.5admins.com if you want to send any questions or your feedback. You can find me at jarrest.com/slash mastodon.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“Unfortunately, apt fetches packages over plain, in the clear HTTP, not HTTPS. And that's something I've complained about to Debian devs before. And I got poo-poohed because they're like, well, we do signature checking, so it doesn't matter. And this bug is an example of why it really does matter… the avenue that we do leave open if we use HTTP transport instead of HTTPS is attackers can still monkey with the actual communication channel itself.” — Jim Salter, 2.5 Admins

Asked about a recent bug in apt, Jim explains that a man-in-the-middle could exploit how apt handled downloads after an HTTP redirect, potentially running code as root. He argues signature checks alone do not make plain HTTP safe.

2.5 Admins · 2026-11-05 Permalink → Listen →
2.5 Admins Around 03:38 into the episode
Jim Salter

You need to sound a little more Canadian.

Joe Ressington

There was a recent bug in apt in Debian, which shone a light on something that you've been talking about for a while, Jim.

Jim Salter

Yeah, unfortunately, apt fetches packages over plain in the clear HTTP, not HTTPS. And that's actually something that I've complained about to Debian devs before. And I got poo-pooed because they're like, well, we do signature checking, so it doesn't matter. And this bug is an example of why it really does matter. Because, okay, sure, you can check the package against a key, although I'm also a little upset that you download the key over plain text HTTP as well. But, you know, if we say, well, all right, we're sure that the system is going to only be trying to use the right key to blah, blah, blah. All right, whatever. But the avenue that we do leave open if we use HTTP transport instead of HTTPS is attackers can still monkey with the actual communication channel itself. And that's what happened in this case. You know, when you man in the middle the HTTP stream, you can try to target hard links on the system in ways that the package manager really wasn't expecting. And you can potentially run code that you really didn't want to run as root as a result of this. If the wrong attacker gets man in the middle on an unpatched version, havoc can ensue.

Joe Ressington

Yeah, but this has been patched at this point. But that doesn't take away the fact that it's still not using SSL to download the packages in the first place.

Jim Salter

I would like to point out that Canonical has had to backport that patch all the way back to 1404 for Ubuntu Pro users. So I really don't want to hear about it's okay because we patched it. Sure, you patched it after how many years of vulnerability out in the wild? We have no way of knowing whether…

Speaker 4

It's 2026, not 96. I can't believe they did this. This is insane.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“If there's one thing reading something like a decade's worth of Backblaze's failure reports has taught me, it's that Backblaze buys as much of whatever they can that's as cheap as possible.” — Jim Salter, 2.5 Admins

The hosts were going through Backblaze's latest hard-drive failure report, the dataset most people reach for when they want to know which drives to buy. Salter's argument is that the numbers describe Backblaze's purchasing strategy more than they describe drive quality: the company buys cheap because failure is expected and its infrastructure absorbs it. He finishes the thought by saying that if you aren't willing to do serious analysis on those numbers, you probably shouldn't be reading them at all.

2.5 Admins · 2026-09-03 Permalink → Listen →
2.5 Admins Around 06:38 into the episode
Joe Ressington

I'm going to have to look now. I don't think I want to, though.

Allan Jude

…buying them. I wonder why. Is it because maybe they were less expensive, even though they had a slightly higher failure rate? Or were they just more available? Or, you know, there was some reason why Backblaze kept buying them anyway. And, you know, it doesn't mean that the Toshiba drives are bad.

Jim Salter

If there's one thing reading something like a decade's worth of Backblaze's failure reports has taught me, it's that Backblaze buys as much of whatever they can that's as cheap as possible. And that's pretty much the criteria. To be fair, you know, they are a bulk storage company, and it absolutely makes sense for them to look at hard drives as something that not might fail and we'd like to prevent that, but will fail. And we just need to design policies that catch those failures before they become larger problems. So what makes sense for Backblaze may not make sense for you. And if you aren't willing to do some serious crunching and analysis when you're looking at Backblaze's numbers, maybe you shouldn't be looking at Backblaze's numbers because they frequently don't mean what you think they do.

Allan Jude

Yeah, if you have the infrastructure to keep replacing the drives when they fail, a drive that fails more but costs less is an easy trade-off to make. If each time you have to buy a drive or deal with a warranty, it's an issue for you, then you might be willing to pay a little bit more for drives…

Jim Salter

And I think it makes sense that we're seeing much lower failure rates in the much larger drives, not only because they're newer, so they haven't had a chance to get to the crappy back side of the bathtub curve, but also keeping in mind, again, that Backblaze buys cheap drives. They want to get the…

Joe Ressington

Changing language for storage capacity. This is a short piece from Michael Lucas, who both of you know.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“…By comparison, modern Ubiquiti access points, while they do have a standalone mode, they are dumber than the absolute stupidest Linksys router from the year 2000 if they're running standalone mode without the controller.” — Jim Salter, 2.5 Admins

The show is on hardware Wi-Fi controllers, which Salter notes cost about as much as one or two access points. He installs both brands at client sites, and his objection to the market leader is not price but what the gear degrades into once the controller is gone. He follows it with the first-generation CloudKey, whose non-journaling filesystem corrupted after about three unplanned power cuts in his testing.

2.5 Admins · 2026-08-20 Permalink → Listen →
2.5 Admins Around 25:24 into the episode
Jim Salter

…you know, these hardware controllers are not at all expensive. The hardware Omada AP controller that I have at a school client of mine is currently running, I think, 140 or so access points. And that was the expensive Omada controller. And I want to say it cost like $200. The cheap ones like... 50.

Joe Ressington

Yeah, the cheap one is roughly the price of an access point, essentially.

Jim Salter

…local and on-premise control, if you want the cloud stuff, Omada can do that as well. It's just an option, not something that you're forced into. And even the controllers themselves are optional, I'd like to point out, because all of the features of the Omada access points are fully accessible when they're run individually as standalone devices, each with its own fully functional web interface. By comparison, modern Ubiquiti access points, while they do have a standalone mode, they are dumber than the absolute stupidest Linksys router from the year 2000 if they're running standalone mode without the controller.

Joe Ressington

Right. Well, we better get out of here then. Remember, show at 2.5admins.com if you want to send any questions or your feedback. You can find me at joeress.com slash mastodon.

Jim Salter

You can find me at mercenary sysadmin.com.

Allan Jude

And alanjude.com for me.

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.

Collections They Appear In