text stringlengths 9 408 |
|---|
[459.72 → 465.94] So even for services that don't have 2FA natively, I can use Thalia to add |
[465.94 → 472.74] 2FA to those services as well. And then just transparently, that token lasts for the length |
[472.74 → 477.08] and duration of your session. You can figure how long it is. I think it's five minutes by default. |
[477.62 → 482.94] Any other service that has the same rule on your system, it will reuse that authentication token. |
[482.94 → 486.14] So if you're switching between a bunch of different services on your system, |
[486.92 → 490.50] log in once, and you're logged into all of them all at once automatically. |
[491.10 → 494.10] Boy, that does seem nice. It has a lot of great features in here. |
[494.72 → 499.20] All right. Well, it sounds like something to check out. We talk about this kind of stuff. Also, |
[499.40 → 504.00] when we talk about things like hosting your own password manager or hosting your own email, |
[504.26 → 507.72] I wonder if there's an audience and I wonder if you out there too, I wonder if there's a portion of |
[507.72 → 514.74] the audience that thinks, I'd never want to host this. This crosses a line. These breaks and things |
[514.74 → 520.52] get hosed. And so I wonder if there isn't just a discussion to be had there, Alex, about where |
[520.52 → 524.28] that line is for us. And if we ever consider dropping something and saying, you know what, |
[524.40 → 526.20] I'll just go with a commercial service for this. |
[526.20 → 533.16] Is this grumpy old man, Alex, coming out here? But my tolerance for stuff breaking, |
[533.94 → 541.26] be it an update, be it a dodgy power cable, whatever the reason, my tolerance is so low these days. |
[541.62 → 546.48] And a couple of episodes ago, I mentioned I had my server on the desk next to me because I was |
[546.48 → 551.60] troubleshooting some issues with hard drives failing. Turns out it wasn't the hard drives failing. |
[551.60 → 557.76] And I wasted maybe three weeks on trying all sorts of different drives and cables and disc controllers. |
[558.22 → 563.90] No, it was the power cable from the power supply to the drives. No kidding. I wasted three weeks on |
[563.90 → 569.38] that damn thing. And that just left me with a huge bitter taste in my mouth. And, you know, |
[569.40 → 575.76] I'm not going to stop self-hosting. Don't, don't worry, audience. We're okay. But there were just |
[575.76 → 581.16] moments in that three-week period. I was like, screw this. I just want to throw the towel in and just be |
[581.16 → 586.24] done and just pay Netflix or whoever it is. Wow. You know, in that same period of time, |
[586.28 → 590.18] I think you also had a certain Home Assistant update not go so well, too. So you were really |
[590.18 → 594.76] getting it from both ends. Yeah, that's the thing. And, you know, this week, another example is |
[594.76 → 601.08] Nextcloud just stopped working for some reason. I have no idea why. I've been sat on Nextcloud 19 |
[601.08 → 607.98] for the last year plus. I haven't touched it. Furthermore, I pull in the regular updates that they provide |
[607.98 → 616.08] to the 19 branch through Docker Hub tagging. But I don't update major versions on Nextcloud unless I |
[616.08 → 621.38] have a good reason to because I don't want all the new stuff. I just use it for documents. That's |
[621.38 → 625.48] pretty much it. Right. I actually figured this out because I got a notification on my phone |
[625.48 → 630.26] because photo upload stopped working. It said this server is in maintenance mode. And I'm like, |
[630.64 → 637.34] OK, yeah, I remember. I've had this before. So you Google it. First result says go into the config file |
[637.34 → 644.26] and edit a single line that says change maintenance mode from true to false in the config.php file. |
[645.42 → 651.62] So I did that. It was already set to maintenance mode false. So I'm like, OK, well, |
[651.62 → 656.64] clearly something's going on here. So I then start looking into the Nextcloud container logs. |
[657.40 → 666.12] And then it said something about character encoding, UTF-8 char sets. And I just at that point, |
[666.16 → 671.30] I was like, oh, no, here we go. This is going to be a long evening. Long story short, I ended up |
[671.30 → 679.40] having to go into the MySQL container itself with a Docker exec command, get into the MySQL nitty-gritty |
[679.40 → 687.52] details, delete and change a couple of data types from Var char to text fields and all this, you know, |
[688.12 → 694.18] stuff that's way over the head of most people doing self-hosting unless this is your day job. |
[695.04 → 700.44] And I just thought to myself, who is this stuff targeted at? If I was paying for Nextcloud as a |
[700.44 → 707.72] hosted service, I mean, these things exist. Like Google Drive is similar. OneDrive, you know, |
[707.72 → 713.24] all these different cloud file repositories. If these things broke, and I had to jump through all |
[713.24 → 718.94] these different hoops to fix them, I would just migrate to a different service. And yet, because |
[718.94 → 726.54] I feel a certain sense of obligation, partly because of this show, but also partly because of my beliefs |
[726.54 → 733.94] towards self-hosting in general, I power through it. But sometimes I just want to throw the towel in, |
[733.94 → 738.24] you know. Did you ever get to the root of why that Nextcloud issue just cropped up? |
[738.90 → 743.26] I think it was because of the character encoding. So what I ended up doing as part of the |
[743.26 → 746.64] troubleshooting process was like, right, okay, well, I'm in the weeds here. I may as well go to |
[746.64 → 748.22] Nextcloud 22 whilst I'm here. |
[748.66 → 750.48] Yeah, that would be my line of thinking too. |
[751.38 → 756.82] Yeah. And then I thought it didn't start because I thought, oh, some database schema changes have |
[756.82 → 763.30] probably happened, and it's just not tested properly on 19. Okay, fine. I'll go to the next major release. |
[763.30 → 769.66] That's stable. Yeah. And so like I said to you, I had to drop to the MySQL command line, |
[769.96 → 776.14] delete a couple of tables that were failing the repair because there's an OCC command. So you can |
[776.14 → 783.94] log into the Nextcloud container and run, there's like a shell, like a CLI tool built into it called |
[783.94 → 790.42] OCC. OwnCloud Client, I would imagine is what it stands for. Don't quote me on that. And you can do a |
[790.42 → 793.66] bunch of maintenance stuff, like you can turn maintenance mode on and off, you can repair |
[793.66 → 797.96] things, et cetera, et cetera. And it kept failing on this line item to do with news articles, |
[797.96 → 802.42] which is a plugin I don't even use anymore. I think I enabled it for testing and then forgot |
[802.42 → 808.74] about it. And much like my Home Assistant update, I probably missed it in the logs because I don't |
[808.74 → 815.18] necessarily look at the logs every day because I've got other stuff to do. And yeah, once I deleted the |
[815.18 → 820.44] news, once I'd gone into the correct database table and then deleted the three or four lines |
[820.44 → 825.92] that were causing the schema migrations to fail, then the upgrade proceeded and it was all fine. |
[826.04 → 827.54] But it took me a couple of hours. |
[828.00 → 833.46] I've had those moments where I thought maybe like a server was dead and gone. And I thought, |
[833.56 → 838.96] I don't, I don't know if I, I don't know if I ever want to rebuild this. And then thankfully |
[838.96 → 841.78] they came back, and then I thought, okay, I better get my backup game. |
[841.78 → 846.86] I better get my backup figured out. Yeah. Because I mean, you really got it. I mean, |
[846.88 → 850.74] you got a hardware failure in your server. You got a software failure in Nextcloud, |
[851.14 → 856.38] a Home Assistant, a pretty dramatic Home Assistant failure based on some deprecation and changes. |
[856.80 → 863.38] I mean, you got it one after another. I recently also had a Home Assistant update go south when the |
[863.38 → 868.82] OS update didn't complete correctly. The system locked up and then rebooted. And then it couldn't |
[868.82 → 874.08] get an IP until I had fiddled with it for quite a while. And for a brief moment, I thought to myself, |
[874.08 → 879.52] should I just bail on this and maybe just go back to a traditional server and Home Assistant core |
[879.52 → 884.34] container and just keep it simple. And then it gets working again. And then I think, okay, no, |
[884.40 → 889.42] this is better. And I like this. And I think to myself in the grand scheme of things, this is the |
[889.42 → 895.92] cost of having full control over these services. And, and it does give me a better appreciation of the |
[895.92 → 902.64] value that the cloud providers are offering. And it makes me also understand how valuable my data |
[902.64 → 908.10] must be if they can pay for these services based on mining my information. Like, wow, there must be |
[908.10 → 913.76] some real value there because it's not cheap. And to keep something reliable like file services for, |
[914.18 → 918.04] you know, a 10-year streak actually takes an incredible amount of effort. |
[918.38 → 922.40] Yeah, you're not wrong. I mean, these, these guys, they're our entire jobs, you know, |
[922.40 → 928.18] site reliability engineers. That's the name of our, you know, supporters club that we have. And |
[928.18 → 935.26] there's an entire job dedicated to keeping things up and stable and reliable. And these guys, |
[935.36 → 940.30] they write software, they write scripts, they do all sorts of ninja magic to keep these services online. |
[940.88 → 945.50] I think part of it, Alex, is giving yourself, and you've done this, you know, the ability to rebuild |
[945.50 → 950.66] fairly quickly. Like you do that. You make, you make, if you adopt something, you make sure that |
[950.66 → 957.84] you can redeploy it as necessary. And those kinds of things, I think they help a lot. But I think |
[957.84 → 962.70] this show should ask more often, is there a cloud service that could do this? |
[963.16 → 968.00] And if there is, why are we not choosing to use it? Because it is a big trade-off every time. |
[968.34 → 972.66] And the more critical the service is to you, I think the more serious you have to consider that. |
[972.98 → 975.72] I'm trying to think of a good example. I suppose wiki would be a decent one. |
[975.72 → 981.44] Because you can probably throw some stuff up on GitHub in a random repository, |
[981.76 → 990.98] a bunch of text files, and call that a wiki if you want to. You could also go to some random website |
[990.98 → 996.96] and buy probably wiki capabilities. But there's nothing quite like hosting your own and keeping |
[996.96 → 1005.12] that kind of information in-house. Because the role a wiki in particular fulfills often is to store |
[1005.12 → 1011.52] proprietary, you know, custom information, you know, maybe Wi-Fi passwords for your house or |
[1011.52 → 1016.42] something, or the code to the safe so that your wife can open it when you're out or something. I don't |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.