GC
Gilles Chehade

Gilles Chehade

Things Gilles Says on Podcasts

Where to Find Them

Gilles Chehade has been a guest on BSD Now , DevOps and Docker Talk: Cloud Native Interviews and Tooling and Arrested DevOps .

Recently: “Modern Backups with Gilles Chehade” on Arrested DevOps (November 2026); “Backup S3, Google Drive, iCloud, Notion with Plakar” on DevOps and Docker Talk: Cloud Native Interviews and Tooling (March 2026); “3: MX with TTX” on BSD Now (September 2013).

What They Said

“The restore, and validating that the restore would work, is much more painful than the actual backup. It's easy to trigger the backup and be done. But making sure it's been tried and restored costs so much more that it's usually not done. In many meetups, if I ask how many people check their backups, if there's two hands raised in the room, that's a big thing.” — Gilles Chehade, Arrested DevOps

Asked how people should approach data protection, Gilles Chehade contrasts infrastructure, which has inventories and reconciliation loops, with backups, which usually run on a cron job and a report someone has to read. Restore testing is the neglected part.

Arrested DevOps · 2026-11-11 Permalink → Listen →
Arrested DevOps Around 11:18 into the episode
Gilles Chehade

…like before running an AI and having a plan in case it is deleted something. Like you move the threat to an actual expectation. Like it's going to happen. And so your only way out is I need to be sure that I can recover when it happens and no longer in case it would happen in the next two years.

Matt Stratton

…okay, I'm going to install some appliance in my, you know, in my data center. You know, I have lots of different ways to think about lots of different requirements. So, like, what are the ways people can be reasoning about what solutions do we have out there? How should we be thinking about this?

Gilles Chehade

So, I think there's two ways to reason about backups. More than backups, it's data resilience in a sense. The first one is that people usually do not categorize their data sets. So, it's more of a, yeah, I have machines, I will back them up, and they're all backed up with the same kind of level. So, this, this, like, in many cases, I work. This caused a lot of issues because the amount of data is huge. If they are not categorized, you spend as much to backup everything. And so, the cost rises, and you actually can't back up as often as you should the most sensitive stuff because you're also capturing the stuff that's less sensitive. So, the first thing is categorization. It's actually sitting down with an inventory of all your data sets and putting a weight in front of each of them with regard to how critical they are and how much effort you should pour into making sure that you can restore them in a certain timeframe. So, some data, if you lose one hour, it's not the same issue as others, where you actually can't use 10 million COVID. So, that's the first thing. The second thing is like an observation that when it comes to infrastructure and…

Matt Stratton

That's what I was just going to say. It seems like, you know, and I'm a big believer in infra-code policy code. I mean, hell, the app that I use for tracking my workshops is work, or not workshops. My workouts for my lifting is a thing called Liftasor. And it's like workouts is code, right? You…

Gilles Chehade

So if you don't work out, you break a CI?

Matt Stratton

Well, no. See? Yeah, there's a feature request. No, it's just I create like you write in code kind of the like your progressions and everything like that. But anyway, the point is, so yeah, but but but backups, right, not really being thought about that way and trustable and having those…

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“A big part of the cloud era is that we have an assumption that cloud providers store backups. That the bucket is a backup. It's not the case. And it's not uncommon that you actually lose something there. It's just that there are so many users that the stats are biased: it happens to no one. But when it happens to you, you're going to be very sorry if you don't have copies.” — Gilles Chehade, Arrested DevOps

Matt Stratton asked for the one thing listeners should remember about data protection. Gilles Chehade says to keep backups whatever tool you use, and says he has seen companies close after losing the only copy of their most critical data.

Arrested DevOps · 2026-11-11 Permalink → Listen →
Arrested DevOps Around 21:21 into the episode
Gilles Chehade

Yeah, but we have one for Microsoft Active Directory. Active

Matt Stratton

…and say, like, okay, cement this and you know, tattoo it on the inside of your eyeballs about thinking about your data protection. What's what's the thing that that people should should go? You know, first thing is go check out Plaque R, but you know, for sure. Yeah, of course. Secondly, after you

Gilles Chehade

secondly, is really to take this into consideration that you should have like, regardless of PlaqueR or anything else, you should have your backups because I have a lot of horror stories. I'm not sure here, but I've seen companies close because they lost the wrong copy. They had no copy for the most critical part. That has happened. And that it happened to a lot of people around me. And I'm just one guy. So I'm sure if you ask around you, you will find people that will tell you they lost something. And a big part of the cloud era is that we have an assumption that cloud providers store backups. The trusty buckets is a backup. It's not the case. And it's not uncommon that you actually lose something there. It's just that there are so many users that you see the stats are being biased. Like it happens to no one. Yeah, but when it happens to you, you're going to be very sorry if you don't have copies. So you should actually do an assert assessment of your data, be it your pictures on Google Photo or your personal bucket and stuff. And just wonder what would happen if I lost that. How depressed would I be? And that should be enough to tell you, okay, I should…

Matt Stratton

I feel like that's sort of the key. And so much in the consumerification where that comes in and people think about, and just even at the consumer level, and even us as technologists probably misunderstanding. I mean, I had a conversation with my ex the other day. She's like, I don't understand why…

Gilles Chehade

a false sense of security and a sense of not losing your data. But people who have lost them have been feeling very sorry about this because they write it. They're right. We are not going to back up your data. Share responsibility. You have to do your backups because we can't necessarily restore.…

Matt Stratton

Fantastic. So everyone, if you head over to arresteddevops.com/slash modern backups, that will have this episode show notes with lots of helpful links. And if you visit arresteddevops.com/slash iTunes, you can leave us a review in the Apple Podcast Store. That helps other people find the podcast. I…

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“Before, you would make the entrance cost to your data elevated by adding layers of protection between the outside world and your internal data. But now the agent is running internally. So you get the same threats, but from inside, and you have to guard against it as if one of your users would be the one deleting your data, which is a very, very different threat model.” — Gilles Chehade, Arrested DevOps

Gilles Chehade lists what changed since his early sysadmin days: ransomware becoming routine, and AI agents getting broad access to infrastructure and data. He had just mentioned a public post by an AI security lead saying an agent wiped her mailbox while she was at lunch.

Arrested DevOps · 2026-11-11 Permalink → Listen →
Arrested DevOps Around 06:17 into the episode
Gilles Chehade

…because backups is not absolutely not a sexy area. No one wants to deal with that. Usually, when you deal with that, it's because you've been tasked to deal with that in a hurry, let's say, and you try to get rid of your task as fast as possible and hope that you never have to look at it again.

Matt Stratton

…of a, for lack of a better word, user-level problem, right? You know, like we're like, oh man, you know, we did this wrong. We have to, we have to put things back to the way that they were. There's sort of like restoring a system because it went south, and then there's rewind time, you know,

Gilles Chehade

right? There's a well, that's a fun anecdote. I'm not going to blame a company because anyway, it was a long time ago. But when I was when I was a student and doing the freelance gigs, I used to work for many small companies where the basically sysadmin is the only techie guy that's in the company. You're the whole tech department. Yeah. Yeah. And the guy is supposedly not supposed to touch this. So he's going to give access to whoever requests it. And if there's a director of something who's absolutely not tech, he's going to have full access to the things. And in many cases, I would go and restore stuff that were accidentally deleted because someone just typed, someone was not supposed to touch that data, actually typed a command that he should not have typed and wiped the entire set of data. And yeah, it's a mix of hardware failure of human disaster skills, let's say. But what's been turned into an interesting situation for me is how the world evolved in the last few years with mainly AI and ransomware. I'll do it in other ways talking about ransomware first, because when I was younger, there was a trend about the crypto crypto vari, which was a…

Matt Stratton

So yeah, it's that threat landscape has changed substantially, right? You know, we think about different things and it seems like almost every day, you know, there's some story about, oh, I accidentally, you know, my AI agent actually dropped this table in production and things like that. And…

Gilles Chehade

I would agree that you do not solve that with backup, but you have to reason about your backups in a different way. Before you would say it's unlikely that I lose data, because it takes like there's very there's statistics about that, yeah, but there's very low statistics, very low odds that you…

Matt Stratton

So I think we've done a pretty good job of talking about why everything is terrible and everyone should be scared as hell and everything. We're not always here just to tell you that you should just go retire off to a kettle and raise chickens in Montana or something like that. So let's think about…

Speaker names come from our own diarization. The position is estimated from where the line sits in the episode.
“Before, you would say there's very low odds that you will lose hardware this year if you have enough redundancy, and low odds that you're going to be attacked if you raise the bar. But this has all changed drastically now. It's no more 'I have low odds that it's going to happen.' It's just: it's going to happen. So your only way out is to be sure that you can recover when it happens.” — Gilles Chehade, Arrested DevOps

Matt Stratton had said an AI agent dropping a production table isn't something backups solve. Gilles Chehade agrees, but says agents change how you should think about backups. His point: data loss goes from an unlikely event to an expected one.

Arrested DevOps · 2026-11-11 Permalink → Listen →
Arrested DevOps Around 08:48 into the episode
Gilles Chehade

…that was not even the discussion before. And of course, war, which is an active topic for us, since we've been contacted by people that had data in data centers that were struck by missiles. So not the kind of stuff I would hear about a few decades ago. Not that funny, but interesting. Interesting.

Matt Stratton

…have it backed up because it was a POC or I never properly operationalized it or anything. And that's all. That is an entirely other topic we can probably talk about on the show, which is how do you get your AI teams to enable teams to think about productionizing apps? But it happens a lot, right?

Gilles Chehade

I would agree that you do not solve that with backup, but you have to reason about your backups in a different way. Before you would say it's unlikely that I lose data, because it takes like there's very there's statistics about that, yeah, but there's very low statistics, very low odds that you will lose hardware this year if you have enough redundancy. And there's low odds that you're going to be attacked if you raise the bar to that level so that an attack actually costs more than that to succeed. But this is all changed drastically now. So it's no more, I have low odds that it's going to happen. It's just it's going to happen. And so now you do not have any other defense to do backups before like before running an AI and having a plan in case it is deleted something. Like you move the threat to an actual expectation. Like it's going to happen. And so your only way out is I need to be sure that I can recover when it happens and no longer in case it would happen in the next two years.

Matt Stratton

So I think we've done a pretty good job of talking about why everything is terrible and everyone should be scared as hell and everything. We're not always here just to tell you that you should just go retire off to a kettle and raise chickens in Montana or something like that. So let's think about…

Gilles Chehade

So, I think there's two ways to reason about backups. More than backups, it's data resilience in a sense. The first one is that people usually do not categorize their data sets. So, it's more of a, yeah, I have machines, I will back them up, and they're all backed up with the same kind of level.…

Matt Stratton

That's what I was just going to say. It seems like, you know, and I'm a big believer in infra-code policy code. I mean, hell, the app that I use for tracking my workshops is work, or not workshops. My workouts for my lifting is a thing called Liftasor. And it's like workouts is code, right? You…

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

Collections They Appear In