wi 1

Password:

password123

wowzersbrain, Honestly I heard blah blah blah wordpress blah blah and I think, the site was probably compromised. There are Wordpress vulnerabilities posted every other week.

+1 for ncdu. I would run this and see where the space is consumed: ncdu -x /

That -x says don't decend other filesystems/datasets. That keeps it out of /dev and such. Run that for each and every dataset. Or don't use -x and see if it runs across everything okay.

I have just gotten in the habit of using -x to keep it out of /dev and /tmp and other datasets. Maybe that's not needed anymore. I don't know.

If you have been using snapshots then you will need to look there: zfs list -t snapshot -o space

And: zfs list -o space

a little further down they mentioned using ufs

I think ufs gained snapshot ability at some point, I actually don't know since I am not using it. But then ncdu should find all of the used space okay.

Sometimes web server log files get blown up due to the external nature of abuse from the hostile internet. That's another typical place to consume all disk.

The database server depending upon the storage engine might expand the db file and not shrink it. The various engines have ways to re-compact. "vacuum"?? Since it is Wordpress it is possible something compromised it and then used a lot of database storage.

Hello, I need help. I added a new geli encrypted mirror to my zroot and now freebsd won't boot anymore

zro_read error: 5

I don't think the bootloader is opening the new geli drives

I set them up with -g

how can i recovery my system to be bootable again?

oh that's so weird

it uh... booted up anyways after lots of errors

ohno

\o

i'm trying to moun a ext2 partition. I get `mount: /dev/da2s2: No such file or directory`, but if I do `file -s /dev/da2s2` I get this:

`/dev/da2s2: Linux rev 1.0 ext4 filesystem data, UUID=eb61a36e-4d8c-4320-beb1-81c18f7695f7, volume name "rootfs" (needs journal recovery) (extents) (large files)`

afaik the journal recovery should happen automatically if I try to mount it

yourfate, try specifying the mounting point and filesystem type (-t).

I do specify the mountpint

what is the correct option for the `-t` flag? just ext2?

yourfate, sorry, I didn't notice it in your message. I get exactly this error if I pass only the device to mount, that is not it fstab.

I'm doing this `doas mount /dev/da2s2 ./mount-linux`

mount linux exists

yourfate, I cannot tell. I read the mount man page yesterday, and saw no explanation.

If you pass the wrong (unsupported) -t, the error message says so.

`doas mount -t ext2 /dev/da2s2 ./mount-linux` returns `mount: /dev/da2s2: Invalid fstype: Invalid argument`

do I need some pacakge? I didn't find anything in the freebsd article on those fses

yourfate, ext2fs manual says -t ext2fs <man.freebsd.org/cgi/man.cgi?query=ext2fs&sektion=5>

I tried that too

then I get invlaid argument

The same man-page says you must have: kldload ext2fs .

I'm installing e2fsprogs rn

oooh

Yes: error messages from `mounm

Error message from `mount' are not very good.

`kldload: can't load ext2fs: module already loaded or in kernel`

hm

Can you mount any other ext2fs system?

I have never tried

I have none

I just have this SD card that was in a pi I was given, and I want to snoop around :P

Try creating one, perhaps, just for testing purposes...

I for one failed to found an NTFS partiotion from the Ventory USB stick. Still no idea how to do it...

I have mounted NTFS and BTRFS successfully before

How did you mount NTFS on FreeBSD?

I tried it with the ntfs-3g utility from FuseFs, but to no avail.

ye, I used that

oooh I fixed it!

How?

I needed the `e2fsprogs`

then I had to fix the journal

using `doas e2fsck /dev/da2s2`

So your initial conjecture was correct, lucky you (-:

then I could mount it using `doas mount -t ext2fs /dev/da2s2 ./mount-linux`

this PI was used in a digital signeage application

it seems they used teamviewer :D

lol

yourfate, You error "/dev/da2s2: No such file or directory" above is misleading. Could that be a bug in `mount' or one of its subprograms?

oof, dd block size can make a huge difference

I'm backing up that old sd card

no bs specified: 1 MB/s

bs=64k: 64MB/s

t'was always so, back to the 1980's

:)

so far I have had dd work fast without BS before

that was with internal drives tho mostly

Local disks are going to have file system buffer cache. SD cards vary greatly in quality and many are awful. Specifying a dd block size is going to make a difference. I usually use 1M at least these days since RAM is available for it.

rwp, what do you mean? why does that make a difference?

Re: 07:17 <yourfate> oof, dd block size can make a huge difference; 07:17 <yourfate> no bs specified: 1 MB/s ; 07:17 <yourfate> bs=64k: 64MB/s

yourfate, try using XFS instead of ext on flash storage. it makes a huge speed difference

rwp, no i mean why does the block size make that big of a difference

how is the qualitiy of an sdcard determined?

Some SD cards are just really crappy. To read a default 512 byte block internally it might read an 8K stripe. Then to read the next 512 byte block it will read the 8K stripe again.

And so a read-amplification happens causing a default block size read of 512 byte blocks to result in 16 reads. A 16x performance penalty immediately.

so if the block size is 8K, why would 1M be any faster then bs=8K?

Which can be mitigated by using a larger block size.

and to be clear, the problem is the sdcard not reporting it's true sector size to the operating system?

As to why 1M might be faster than 8K I imagine it is just all of the overhead. Doing the read might set up a DMA and stream most of that out all at once. Which is just more efficient. And some Flash NAND storage might have even bigger stripes than 8K.

AFAIK dd does not ask the storage for the logical block size. It just does a read(2) using the block size we tell it.

And the default block size unless we tell it something different is 512 bytes.

I've heard of people using pv instead of dd. Does pc read the block size

I use pv and like the progress bar. It also does not read the block size afaik.

I am usually using pv to read from local disk piping to dd to write to an NAND Flash though.

pv foo.img | dd of=/dev/da0 bs=1M iflag=fullblock oflag=sync,direct

filesystems read block sizes though right?

just not raw-reads

That way I get a nice pv created write progress bar. And dd writes to my usb connected storage /dev/da0 with a bigger block size.

When file systems are created I believe they do read the block size. But flash storage is natorious for lying. That's why there is the flashbench program to deduce it from benchmarks.

oh it's not in ports?

Not that I see.

But honestly I would have referenced the URL of the upstream just as documentation for looking at what it does.

thankyou for this tool. this is probably why XFS is so much faster

XFS is an awesome file system. +1

is XFS well supported on FreeBSD too?

i know it is on Linux

I don't know. I haven't needed to try it.

dev.c:21:10: fatal error: 'linux/fs.h' file not found

21 | #include <linux/fs.h>

damn, flashbench on bsd is not to be

On FreeBSD I am one of the ZFS converts.

===> NOTICE:

The xfsprogs port currently does not have a maintainer. As a result, it is

I see that "currently does not have a maintainer" for a great many of the ports. So... Have become desensitized to it.

But I did not know that flashbench was tied that deep with the Linux compile environment. Oh well!

sfox, forums.freebsd.org/threads/xfs-support.61449 says xfs is supported through fusefs and shows it working.

I have no idea how well that works though.

i like ZFS too bug XFS is still faster for some reason

i like ZFS too but XFS is still faster for some reason

I think it was not terribly long ago on 2.5 Admins that they talked about using pv instead of dd, and that it was often faster without having to specify blocksize

the pv man says: "The default buffer size is the block size of the input file's filesystem multiplied by 32 (512kb max), or 400kb if the block size cannot be determined."

Hello, all. Does anybody have experience with extracting the EDID data from your laptops? On my HP 15-bs0xx I get this: <pastebin.com/raw/yPta5bHa>

I'm using bsd.prog.mk but I don't seem to find an option to make it move object files to a custom directory

sfox: oh, I'm not using ext by choice, its just an old SD card I found and watned to see whats on it

im uing UFS on the sd card