Following or sort of following instructions here, which I found a link to from here. Well, actually didn’t click the source link, until I couldn’t figure out how to link to the post on the second link, the comment with those instructions.

My sparsebundle isn’t using HFS. I’m using Disk Utility first aid again.

I also mounted the network share with Finder. I actually made a new Time Machine backup, and deleted the corrupted file, but that drive is backed up, so I just copied the corrupted files to the Raspberry Pi, so my computer can be rebooted or crash or anything, and it won’t stop the Mac mini from trying to repair it.

I didn’t run one of the first commands the first time I tried fixing it, the chflags command.

Also, that network share, doesn’t have the Time Machine flags in Samba. You have to make it’s own share folder if you want that. I just want to repair the image. I can copy it elsewhere if I want to use it.

How do I copy the new backups to the old backup file?

Don’t waste time with disk recovery software, at least Disk Drill or whatever, it found no files. If you mount it, there appears to be data, yet Disk Drill can’t find anything. Didn’t try restoring anything.

I was originally trying some other recovery software, but decided I didn’t want a yearly program, no idea if it auto renews, or you just get updates for a year. You can pay $29 extra for Disk Drill, and get updates forever. Still a lot of money. But I think it can recover data from EXT4. So it could be useful in the future, if multiple drives crap out at the same time.

Finder won’t let you copy anything from the sparsebundle. As that is what I was going to do, copy to a new non corrupt image. Can’t delete the corrupt stuff either. Linux might be able to read it, but not write. Yes, I want my old backups. Somebody should make a Linux program to recover it, I’ll pay for it, if it’s cheaper then disk recovery software.

I wonder if this lets you work with sparsebundle files. Or maybe you use the other thing, but instead of read only driver, use that. Too bad, my kernel is too new. So I won’t buy it. Yes, such a program, would also need to work with the latest Linux kernel, and continue as I update my computer. My kernel is 6.1.8-1-default. Using VMs is too much work.

Too bad macOS can’t fix stuff it creates.

One data recovery program is I think $90 per month. You can probably cancel it, but that’s still a rip off. Doesn’t say what the lifetime version is. Do you pay that every life? Every year? Every 100 years?

I needed to update the Mac mini, that’s why I created a new Time Machine backup. And it’s funny how somebody thinks USB is more reliable for Time Machine, then Ethernet. Servers use Ethernet, and have no issues usually. So USB is less reliable in my opinion. Now Ethernet + USB, that’s as unreliable as USB directly. But probably not less reliable.

Update
Couldn’t save it, deleted it from Raspberry Pi, well not the Pi’s backup drive, if it’s backing up with rsync, it might have the “–delete” option. But rsync backups is a bad idea, if the source drive is failing and corrupted data is being replaced. So a backup of the Time Machine image, wouldn’t have saved it, unless I noticed soon enough.

I could change to a better backup method but I’m lazy. Well, wouldn’t change from Time Machine on the Mac, for the Linux computer. Perhaps I will for my pics folder. Problem is, it keeps multiple days, which may require more space.

Perhaps check dmesg for USB errors on each drive before running rsync, so the backup isn’t possibly destroyed.

Either one backup drive is dead, or the cable is. My Time Machine backup for Mac mini was on it. I copied the it to a different, 1 TB smaller drive, and now it’s running fsck on the backup image. But I will need to resize the partition, because it won’t think there’s any free space if I don’t.

I assume fsck is still running, not on Mac mini right now.

So I got a shitty WD portable drive from Best Buy, Target didn’t actually have it, so I had to transfer $200 total from savings, thanks to Target’s authorization. I tried using the cable from the old Seagate drive, because it’s longer, and got some errors, one error being some “critical target error”. Which the other drive has too. The other errors might not be important.

[Sun Jan 29 14:41:05 2023] ses 13:0:0:1: Wrong diagnostic page; asked for 1 got 8 [Sun Jan 29 14:41:05 2023] ses 13:0:0:1: Failed to get diagnostic page 0x1 [Sun Jan 29 14:41:05 2023] ses 13:0:0:1: Failed to bind enclosure -19

Smart works on this drive, without doing any workarounds though. You might have to disable UAS or whatever for the Seagate drive.

smartctl 7.3 2022-02-28 r5338 [x86_64-linux-6.1.8-1-default] (SUSE RPM)
Copyright (C) 2002-22, Bruce Allen, Christian Franke, www.smartmontools.org

Read Device Identity failed: scsi error unsupported field in scsi command

If this is a USB connected device, look at the various --device=TYPE variants
A mandatory SMART command failed: exiting. To continue, add one or more '-T permissive' options.

That’s what happens if I run smartctl -a on the other Seagate drive, and the possibly dying or dead one.

If I didn’t buy the brand of the first drive that died on me, I wouldn’t buy WD. Or maybe it didn’t die on me, it was just dead. Might not have been my computer, a family computer perhaps.

But all regular USB drives are crap most likely. So let’s see how long portable drives last. Well, might last a long time, since Time Machine won’t be on it anymore. Didn’t think it was a big deal only having one copy of Time Machine. I was wrong. There’s two copies now.

I need to test the other USB cable with either the portable drive, or the dead drive. If it too gets errors, then it isn’t the USB cable.

But if it works fine, USB cables apparently die. Too bad it’s copying data from that drive, in case the movies drive dies too. It’s the same model of drive I think as the dead one. Well, I’ll have two backups of the movies drive if the other drive is fine. Now where do you buy a replacement cable? Don’t even know what kind of cable it is.

Then again, I might replace the movies drive with a WD, since SMART works on WD. Not anything to do with reliability. If I wanted that, I would shell out money for NAS drives. But it might have only been $125 for a NAS drive and two bay HD dock from Best Buy, not much more. But I only need one slot in the dock. And would rather have an enclosure, but they have no 3.5" enclosures at the store. Most of the stuff on Amazon is junk.

The last Time Machine backup was from the 26th, but I didn’t notice the problem until the 28th.

And if you shrink the partition, it runs fsck again. I was going to back it up before updating it. Didn’t know there was a new macOS, not the latest latest. But one version behind the latest major version. Won’t be done till tomorrow probably. And might just fail anyways. As fsck said it was corrupt or something and it couldn’t repair it. Oh and it makes a big new partition, hopefully I can delete that partition if it shrinks it. I should have made it 1.5 TB or 2 TB though. 1 TB is kind of small, when the internal drive is around 500 GB I think. At least it won’t use all the space on the drive.

I can mount the image though. So some data must still be in it. Might have to mount and copy to new image. So maybe sometime next week I’ll have backups again on it.

Now you know why USB is unreliable, USB cables can die, but so can SATA cables. Assuming that’s the problem.

Was going to use it with my Mac mini, but I don’t think the paid ext4 driver supports TRIM. There’s a trial available. I uninstalled it, and put the security back on full security on the Mac mini.

If the Mac mini dies, and it’s formatted as APFS, reading it might be challenging. There might be a way, but I’m way too lazy to look it up. The kernel doesn’t have APFS support enabled, so I need a custom kernel? No thanks.

I put a Kingston NV2 2TB M.2 2280 NVMe Internal SSD in it. For USB it should be fine, apparently after you fill it up, the write speed decreases, it’s USB so who cares.

                        KDiskMark (3.1.2): https://github.com/JonMagon/KDiskMark
                    Flexible I/O Tester (fio-3.33): https://github.com/axboe/fio
--------------------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes
[Read]
Sequential   1 MiB (Q=  8, T= 1):   841.750 MB/s [    822.0 IOPS] <  9697.37 us>
Sequential 128 KiB (Q= 32, T= 1):   550.141 MB/s [   4298.0 IOPS] <  7439.93 us>
Random   4 KiB (Q= 32, T=16):    13.509 MB/s [   3378.3 IOPS] < 37084.04 us>
Random   4 KiB (Q=  1, T= 1):    27.277 MB/s [   6819.3 IOPS] <   145.97 us>
[Write]
Sequential   1 MiB (Q=  8, T= 1):   874.840 MB/s [    854.3 IOPS] <  8992.19 us>
Sequential 128 KiB (Q= 32, T= 1):   567.175 MB/s [   4431.1 IOPS] <  7164.86 us>
Random   4 KiB (Q= 32, T=16):    16.348 MB/s [   4088.2 IOPS] < 30577.87 us>
Random   4 KiB (Q=  1, T= 1):    33.294 MB/s [   8323.7 IOPS] <   118.17 us>
Profile: Default
Test: 1 GiB (x5) [Measure: 5 sec / Interval: 5 sec]
Date: 2023-01-09 19:38:54
OS: opensuse-tumbleweed 20230108 [linux 6.1.3-1-default]

And don’t ask me how the write speed is faster. Usually it’s the other way around. That’s plugged into the Type-C port on the back of my motherboard.

To get the above speeds, make sure the drive is connected @ 10000M. When I ejected it, and unplugged it and back in, only 5000M. Had to unplug and plug back in again. lsusb -t will tell you the speed it’s connected at.

I think the highest temperature I got was 40C on Linux desktop, can’t read temperature sensor in macOS. It’s macOS 12, I’m not updating it anytime soon. Another reason to not to use the Mac mini.

It stayed under 35C when copying my pics folder to it. I used the thermal pads, even though a review says you’ll damage it, probably not the SSD I used, it has a sticker over the chips. I’ll ruin the sticker possibly.

Probably didn’t really need it though. I only have my pics and Music on it. Not sure I want it the primary disk for my pics and music. Some say USB isn’t reliable. Would suck if the pics got corrupted, and no good backup existed.

Oh review on enclosure also says the cables it come with are no good. So perhaps a different cable will always connect at the right speed. I probably won’t buy a different cable though. Why bother? What USB speed does my other C cable do? It’s the one for my SATA enclosure.

Now I might buy a new cable, if I reboot and notice it’s slow. But I probably won’t notice, since I won’t run a benchmark, or run lsusb -t.

You don’t want your file system proprietary, unless you have more then one proprietary computer. Well, the iPad might be able to read it. If I need to access all my drives on the Mac mini, I can buy the ext4 driver. Or use the free one, and read only.

In the future, Linux may be ready for M1 Macs. So I can just ditch macOS all together. Or put Linux on that external SSD. Considering I may put nothing else on the drive, it doesn’t matter what I do with it. 1.5 TB free.

The cables are short, luckily it can still sit on the top of my case.

There’s some other enclosure that might let you see the temperature in macOS. It was on Reddit, and I forgot what it’s called. It might be USB4 and Thunderbolt. So faster.

If I had a heatsink, I could have just put this inside my computer, and move the games to it. But the current games drive, has it’s own heatsink, so may not fit in enclosure, and I’m not removing it’s heatsink. Also, the write speed should be fine on the current games drive, if it ever gets completely full. It’s only PCIe 3.0 I think, and the write speed is I think faster then my Crucial PCIe 4.0 SSD, the Crucial has a much faster read speed though. I must not have bought it for the write speed. And mostly read on that drive anyways. Crucial seems to be reliable, the main reason I bought the one I did.

The ADATA SSDs with a heatsink might be reliable enough. That games drive isn’t backed up, the saves for the games are though. Might take a while to back that drive up to the external NVMe SSD. I’d rather use a mechanical drive for backing it up. They are cheaper.

I could fill it up with porn, of myself. With the head replaced, with a woman head generated with Stable Diffusion, from my own head.

I think TRIM works automatically in openSUSE. You apparently shouldn’t use discard mount option.

If I ever run out of space on my internal games drive, I can then add that drive to Steam, and install more games I’ll never play.

Update

                        KDiskMark (3.1.2): https://github.com/JonMagon/KDiskMark
                    Flexible I/O Tester (fio-3.33): https://github.com/axboe/fio
--------------------------------------------------------------------------------
* MB/s = 1,000,000 bytes/s [SATA/600 = 600,000,000 bytes/s]
* KB = 1000 bytes, KiB = 1024 bytes
[Read]
Sequential   1 MiB (Q=  8, T= 1):   899.911 MB/s [    878.8 IOPS] <  9046.62 us>
Sequential 128 KiB (Q= 32, T= 1):   845.219 MB/s [   6603.3 IOPS] <  4819.82 us>
Random   4 KiB (Q= 32, T=16):    12.869 MB/s [   3218.3 IOPS] < 48368.35 us>
Random   4 KiB (Q=  1, T= 1):    27.768 MB/s [   6942.0 IOPS] <   143.39 us>
[Write]
Sequential   1 MiB (Q=  8, T= 1):   869.938 MB/s [    849.5 IOPS] <  9013.61 us>
Sequential 128 KiB (Q= 32, T= 1):   564.985 MB/s [   4413.9 IOPS] <  7176.05 us>
Random   4 KiB (Q= 32, T=16):    16.326 MB/s [   4082.8 IOPS] < 30572.62 us>
Random   4 KiB (Q=  1, T= 1):    33.252 MB/s [   8313.3 IOPS] <   118.15 us>
Profile: Default
Test: 1 GiB (x5) [Measure: 5 sec / Interval: 5 sec]
Date: 2023-01-10 12:32:07
OS: opensuse-tumbleweed 20230109 [linux 6.1.3-1-default]

Can’t tell you why the read speed is better with that benchmark. Don’t think I rebooted the computer. I did update some stuff, and logout, and back in. Perhaps KDE was indexing the drive or something. You use smartctl in Linux to get the temperature, doesn’t work in macOS, even after installing it with Homebrew. macOS thinks it doesn’t support SMART. Good luck fixing macOS.

Nice HTML Hugo or whatever this site is using. It’s messing the JavaScript code highlighter up, and yes, I’m too lazy to use Hugo’s code highlighting, if it does so. Also, way too lazy to edit old posts.

Update #2
If you copy over 70 GB, it gets hot, 54C I think it was when I checked. Running a benchmark to see if it gets hotter. The temperature seems to be going down still. Seems hot for copying @ 130.599 MiB/s. Copying from a 1 TB mechanical drive, so I can delete stuff from it. I know it was hot, cause I touched the drive.

PCIe 4.0 drives run hot, but I think because of the speed, it can’t get the max speeds. So perhaps the enclosure is useless. Remove the top I guess. Need more data to copy to it, apparently a benchmark doesn’t make it that hot. Copying more data, but might be less then 20 GB worth of data. I’m sure I can find more to copy after. Seeing if I can fry the SSD.

Only two degrees more then the hottest my Crucial SSD got. That’s PCIe 4.0 too. And I think it has a heatsink on it.

Still not that hot. And two rclone’s running. Maybe a benchmark too, will get it back up to 54C.

The macOS SSD benchmark I have, runs until you stop it. That’s the easiest way to test it. Too bad no ext4 support in macOS with TRIM support.

I think I might know why it’s getting hot, the SSD might not be making contact with the top of enclosure. Probably pushed it too far down. Where the rubber thing is, that secures it in place. Should probably wait till it’s done copying to fix it. And don’t think that really matters. But it was upside down, that might matter, if it’s sitting on a surface. Under 40C now, and might be cooler then without the top part of enclosure on. The fucking cable won’t let it lay flat on my case.

Only got it up to 41C. Ahh 42C. The two rclones are still running. And 43C, then back down.

Well, I think the drive is fine now, 33C, still copying a bunch of data, mechanical drives are slow.