Rclone, Badblocks, SMR, OVH, Btrfs
Rclone (@ Wikipedia) + FileZilla Server (@ Wikipedia) - FTPS - It's my today's saddening experience. It seems that Rclone and it's FTPS implementation just doesn't work with FileZilla FTPS server and it's implementation. Bleep! But it's so close that it was worth of trying. Yet, many users are suffering, and there's no solution. Several suggestions how to fix, and of course none of those work. Nor the latest -DEV version for Rclone nor FileZilla Server. GggggGGrraaaH. Business as usual, messy, frustrating, enraging and finally just hopeless. - Of course alternate solutions work, but just this combination is broken. magic keywords: TLS session resumption, Explicity FTP over TLS, "TLS session resumption on data connection failed." = Fail Error Problem Kaboom - It means that the data channel should use session resumption with the control channels keys. Instead of creating a new TLS session for data channel. If clients do not implement that, well, you're out of luck then.
Badblocks (@ Wikipedia) Why all code is bad? The progress percent is clearly incorrectly shown, when you specify start and end blocks for the operation. Yawn, always, bad, code. Of course another question is, god forbid, why we have storage devices which require operations like what the badblock does. Well, unfortunately we do.
Wonderful world of SMR (@ Wikipedia) drives. I've got a few drives, that do remap sectors, sure. But only after producing non-recoverable read errors and corrupting data. I though the remapping was designed to happen BEFORE this, when the ECC (@ Wikipedia) indicates that the data quality isn't what it's supposed to be. Well, meh. Horror drives. And of course the health check passes. Users using drives like this suffer from repeated data corruption, and the drive in general looks still good. Which is a very bad combination.
OVH management console and it's dark patterns. There are several ways to cancel service, but only a few of those actually do work. You do all selections, and click ok. But nothing actually happens. This dark pattern is especially enraging, because then they keep charging for cancelled services. Always make additional support ticket when ever canceling anything with OVH and making sure that the service is actually cancelled. They've been playing total asses every freaking time, and getting the things straight later turns out to be very slow and agonizing process. Just be warned, not to never save any payment information with OVH and only use burner payment methods with their services. As well as create a service specific accounts, not putting several services under one account, because that can cause serious problems.
Btrfs (@ Wikipedia) fun, oh yeah. Finally ended up with a full file system, that's nice. Nothing new or special about that, it's quite normal, and a regular thing to encounter. Except when you're using btrfs, it leads to well, rebuilding everything. You can recover the data from the full disk in read-only mode, but you can't free space or mount the disk in read-write mode anymore. Uh oh. Just like with SQL databases, where you'll end up with situation where you can't delete anything, because there's no space left for journal allowing you to delete things. Joy. In some cases this could cause pretty hairy and very hard to recover from situations. - Totally unrecoverable for normal users and extremely painful for experts. - Just copied data off disk, recreated file system, and copied data back. - Not that hard, if you can do it easily.
Stealth Bus Technology. That's quite an amazing feat! It's even officially acknowledged by officials, that around 10% of buses in Helsinki do currently use stealth technology. We have this tracking application called Bus Radar. And busses without proper GPS signal and mobile data connection aren't visible on it. Making the cool real-time information system practically disinformation system, as well as creating a trendy phenomenon of stealth buses being available in the city. kw: häivebussi, stealth bus, HSL
2025-02-09