XFS problems

[Doug Laidlaw]

I have been saving my backups to an XFS partition on an external Seagate
drive. I tried to do it with Ext4, but had too many failures.

For the past month or so, we have been experiencing very severe storms,
some with lightning. After the last one, my external disk became
inaccessible. Looking around the Web, I find that this was a regular
occurrence with CentOS 7, whose default filesystem was XFS. Usually,
running xfs_repair is all that is needed, but this time, no primary or
secondary superblocks can be found, and xfs_repair gives up. The
computer does not now recognize the device.

Are there any other steps I can take? btrfs is suggested as ideal for
backups, but it is largely incompatible with any other filesystem, and
it is said to be overkill for a single workstation. In any case, that
is a question that arises only if I have to buy a new drive.

[Gilberto F da Silva]

Post by Doug Laidlaw
Are there any other steps I can take? btrfs is suggested as ideal for
backups, but it is largely incompatible with any other filesystem, and
it is said to be overkill for a single workstation.  In any case, that
is a question that arises only if I have to buy a new drive.

External discs are fragile devices. I've heard from several people
complaints about it. One of my external discs stopped working suddenly.
I was able to recover some data using
https://www.cgsecurity.org/wiki/testdisk

[Doug Laidlaw]

Post by Gilberto F da Silva

Post by Doug Laidlaw
Are there any other steps I can take? btrfs is suggested as ideal for
backups, but it is largely incompatible with any other filesystem, and
it is said to be overkill for a single workstation.  In any case, that
is a question that arises only if I have to buy a new drive.

External discs are fragile devices. I've heard from several people
complaints about it. One of my external discs stopped working suddenly.
I was able to recover some data using
https://www.cgsecurity.org/wiki/testdisk

Yes, I used testdisk years ago. Regarding an alternative, my computer
has provision for installing 2 inch drives. These drives have been
popular with the Raspberry Pi, because of their small size. I have no
experience of them, but they seem to be solid-state and usable anywhere.

[Gilberto F da Silva]

Post by Doug Laidlaw
I have been saving my backups to an XFS partition on an external Seagate
drive.  I tried to do it with Ext4, but had too many failures.

This does not make sense in my head. Use a different file system tool to
repair your XFS partition.

[Doug Laidlaw]

Post by Gilberto F da Silva

Post by Doug Laidlaw
I have been saving my backups to an XFS partition on an external Seagate
drive.  I tried to do it with Ext4, but had too many failures.

This does not make sense in my head. Use a different file system tool to
repair your XFS partition.

I need to rephrase that. My initial filesystem for this partition was
Ext4. Then I switched to XFS. fsck cannot handle XFS, although it is
suggested that running fsck on it on boot-up is useful. XFS has is own
tool, xfs_repair, but as its very name suggests, it is a tool for
repairs, not for ordinary checks on boot-up. I have used it in the
past, but this time, the damage was too great.

[Bobbie Sellers]

Post by Gilberto F da Silva

Post by Doug Laidlaw
I have been saving my backups to an XFS partition on an external Seagate
drive.  I tried to do it with Ext4, but had too many failures.

This does not make sense in my head. Use a different file system tool to
repair your XFS partition.

I need to rephrase that.  My initial filesystem for this partition was
Ext4.  Then I switched to XFS.  fsck cannot handle XFS, although it is
suggested that running fsck on it on boot-up is useful.  XFS has is own
tool, xfs_repair, but as its very name suggests, it is a tool for
repairs, not for ordinary checks on boot-up.  I have used it in the
past, but this time, the damage was too great.

Regarding the reliability of external disks.
Every use GSmartControl?
I used to determine the age and reliability of internal and External
disks. My external disks are much more reliable than the old disks
(all now replaced but one)on my used machines. And I had no problems
until after the middle of November 2021 with my backups via Timeshift.

I don't know if Mageia has this valuable tool or not.


bliss - brought to you by the power and ease of PCLinuxOS
and a minor case of hypergraphia

[Bit Twister]

Post by Bobbie Sellers
Regarding the reliability of external disks.
Every use GSmartControl?
I used to determine the age and reliability of internal and External
disks. My external disks are much more reliable than the old disks
(all now replaced but one)on my used machines. And I had no problems
until after the middle of November 2021 with my backups via Timeshift.
I don't know if Mageia has this valuable tool or not.

Yep. I have it installed and get occasional messages from my smartd cron job.
journalctl | grep smartd
Jan 01 04:18:07 smartd[761]: Device: /dev/sdb, previous self-test completed without error
Jan 01 10:18:07 smartd[761]: Device: /dev/sdb, SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 21 to 20
Jan 03 05:18:07 smartd[761]: Device: /dev/sdb, SMART Prefailure Attribute: 1 Raw_Read_Error_Rate changed from 109 to 110
Jan 03 05:18:07 smartd[761]: Device: /dev/sdb, SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 20 to 21
Jan 03 09:48:07 smartd[761]: Device: /dev/sdb, SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 21 to 20
Jan 08 03:48:07 smartd[761]: Device: /dev/sda, previous self-test completed without error
Jan 08 03:48:07 smartd[761]: Device: /dev/sdb, self-test in progress, 10% remaining
Jan 08 04:18:07 smartd[761]: Device: /dev/sdb, previous self-test completed without error
Jan 08 13:48:07 smartd[761]: Device: /dev/sdb, SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 20 to 19
Jan 10 01:18:07 smartd[761]: Device: /dev/sdb, SMART Prefailure Attribute: 1 Raw_Read_Error_Rate changed from 110 to 111
Jan 10 01:18:07 smartd[761]: Device: /dev/sdb, SMART Usage Attribute: 195 Hardware_ECC_Recovered changed from 19 to 20

[TJ]

Post by Doug Laidlaw
I have been saving my backups to an XFS partition on an external Seagate
drive.  I tried to do it with Ext4, but had too many failures.
For the past month or so, we have been experiencing very severe storms,
some with lightning.  After the last one, my external disk became
inaccessible.  Looking around the Web, I find that this was a regular
occurrence with CentOS 7, whose default filesystem was XFS.  Usually,
running xfs_repair is all that is needed, but this time, no primary or
secondary superblocks can be found, and xfs_repair gives up.  The
computer does not now recognize the device.
Are there any other steps I can take? btrfs is suggested as ideal for
backups, but it is largely incompatible with any other filesystem, and
it is said to be overkill for a single workstation.  In any case, that
is a question that arises only if I have to buy a new drive.

A backup system that isn't reliable isn't much of a backup system.

If I were in your shoes, even if I could find a way to "repair" the
drive and get it working, I would never be able to trust it again. I'd
buy a new drive. In fact, I'd probably buy two, so I had a backup for my
backup.

Regarding the file system, I use ext4 just about everywhere. Exceptions
are on flash drives or memory cards that are used between Linux and
other systems, where the job requires FAT32, NTFS or EXFAT. I've had a
few failures over the years, but none where I would assign blame to the
file system.

A very few of my failures were hardware-related, heads crashing, bad
cable, lightning, etc. But the vast majority of my failures were my own
damn fault, caused by me doing something stupid when I knew better.

TJ

[faeychild]

Post by TJ
A very few of my failures were hardware-related, heads crashing, bad
cable, lightning, etc. But the vast majority of my failures were my own
damn fault, caused by me doing something stupid when I knew better.

Yes!! I believe that I also run with your pack :-)

Regards

[Gilberto F da Silva]

Post by Doug Laidlaw
I have been saving my backups to an XFS partition on an external Seagate
drive.  I tried to do it with Ext4, but had too many failures.
For the past month or so, we have been experiencing very severe storms,
some with lightning.  After the last one, my external disk became
inaccessible.  Looking around the Web, I find that this was a regular
occurrence with CentOS 7, whose default filesystem was XFS.  Usually,
running xfs_repair is all that is needed, but this time, no primary or
secondary superblocks can be found, and xfs_repair gives up.  The
computer does not now recognize the device.
Are there any other steps I can take? btrfs is suggested as ideal for
backups, but it is largely incompatible with any other filesystem, and
it is said to be overkill for a single workstation.  In any case, that
is a question that arises only if I have to buy a new drive.

Regardless of the size of the disk, we can always fill them. But the
amount of original data is small. An account on a virtual disk such as
GoogleDrive or MEGA can avoid stress of a loss of data. Mageia can have
a folder with automatic synchronization with MEGA.

[Aragorn]

btrfs is suggested as ideal for backups, but it is largely
incompatible with any other filesystem [...

What exactly does that mean, "incompatible with any other filesystem"?

btrfs IS a filesystem, and you create it just as you would create an
ext4, XFS, JFS or reiserfs filesystem.

...] and it is said to be overkill for a single workstation.

It has features that you probably wouldn't use, but that doesn't make
it a bad choice. For most part, it is a self-healing filesystem, it
uses CRC32 checksums, and it is very fast.

I am running btrfs as the filesystem for most of my partitions here,
and have been for close to three years, without any problems. I don't
use subvolumes — I still use traditional (GPT) partitions — but I use
it on both the SSD with all of my system partitions and on the HDD that
holds my Timeshift backups. (They are rsync'ed backups, by the way,
not btrfs snapshots.)

[David W. Hodgins]

Post by Doug Laidlaw
I have been saving my backups to an XFS partition on an external Seagate
drive. I tried to do it with Ext4, but had too many failures.

The biggest problem I've seen in the past with external drives is using an
enclosure that isn't well ventilated, causing the electronics to over heat,
which will result in problems.

Regards, Dave Hodgins

[red floyd]

On Mon, 10 Jan 2022 05:43:59 -0500, Doug Laidlaw

Post by Doug Laidlaw
I have been saving my backups to an XFS partition on an external Seagate
drive.  I tried to do it with Ext4, but had too many failures.

The biggest problem I've seen in the past with external drives is using an
enclosure that isn't well ventilated, causing the electronics to over heat,
which will result in problems.
Regards, Dave Hodgins

I'm seeing that with external NVME enclosures. So I only plug those in
when needed, otherwise they get too frickin' hot.

[David W. Hodgins]

Post by red floyd
I'm seeing that with external NVME enclosures. So I only plug those in
when needed, otherwise they get too frickin' hot.

It happens with spinning rust drives too, if they are left plugged into power for
too long.

Regards, Dave Hodgins

[Doug Laidlaw]

Post by David W. Hodgins

I'm seeing that with external NVME enclosures.  So I only plug those in
when needed, otherwise they get too frickin' hot.

It happens with spinning rust drives too, if they are left plugged into power for
too long.
Regards, Dave Hodgins

Good to see you are still around, Bits. I opted for a "WD Blue"
solid-state drive. Now I am in the usual vicious circle: the device
needs to be initialized to become visible, and it is not visible, so
initialization is impossible. There is a tutorial on YouTube for this
very device, but it assumes that the drive can be found.

I hadn't heard of GsmartControl, but my mobo has SMART available.

[Doug Laidlaw]

Post by Doug Laidlaw

On Mon, 10 Jan 2022 19:34:37 -0500, red floyd

I'm seeing that with external NVME enclosures.  So I only plug those in
when needed, otherwise they get too frickin' hot.

It happens with spinning rust drives too, if they are left plugged into power for
too long.
Regards, Dave Hodgins

Good to see you are still around, Bits. I opted for a "WD Blue"
solid-state drive.  Now I am in the usual vicious circle: the device
needs to be initialized to become visible, and it is not visible, so
initialization is impossible.  There is a tutorial on YouTube for this
very device, but it assumes that the drive can be found.
I hadn't heard of GsmartControl, but my mobo has SMART available.

GsmartControl is in the repos, but it didn't seem to help.

[Bobbie Sellers]

Post by Doug Laidlaw

Post by Doug Laidlaw

On Mon, 10 Jan 2022 19:34:37 -0500, red floyd

I'm seeing that with external NVME enclosures.  So I only plug those in
when needed, otherwise they get too frickin' hot.

It happens with spinning rust drives too, if they are left plugged into power for
too long.
Regards, Dave Hodgins

Good to see you are still around, Bits. I opted for a "WD Blue"
solid-state drive.  Now I am in the usual vicious circle: the device
needs to be initialized to become visible, and it is not visible, so
initialization is impossible.  There is a tutorial on YouTube for this
very device, but it assumes that the drive can be found.
I hadn't heard of GsmartControl, but my mobo has SMART available.

GsmartControl is in the repos, but it didn't seem to help.

GSmart Control is to determine the condition of your drives.

Check your cabling and positioning of the drive.

Have you tired to find it with GPartEd?

Otherwise you hay have bought or drawn a dud drive,

bliss - to tired to even say more.

[Bit Twister]

Post by Doug Laidlaw
Good to see you are still around, Bits. I opted for a "WD Blue"
solid-state drive. Now I am in the usual vicious circle: the device
needs to be initialized to become visible, and it is not visible, so
initialization is impossible.

If it were me I would get/burn/load the latest systemrescue-8.07-amd64.iso
Click gparted in the bottom tray icom and see if gparted can create a disk
partition table, a partition and format it.

https://www.system-rescue.org/
http://www.sysresccd.org/Download has a link to howto burn
iso contents to usb if needed. Do not use any other method
to do so.

If want you can install Mageia gparted and run it as root.
Just make sure to pick the correct device.

[Doug Laidlaw]

Post by Bit Twister
If it were me I would get/burn/load the latest systemrescue-8.07-amd64.iso
Click gparted in the bottom tray icom and see if gparted can create a disk
partition table, a partition and format it.

Did that. I have a recent copy of system rescue. Also tried the one for
Mageia. It can't find a drive either. I was thinking that once I have
a partition table, I would be set, but I can't get that far, even.

I still have Rescatux, but it seems to be pretty useless. Years ago, I
had a rescue disk with all the manufacturer's setup tools, but no
longer. Checking the cables next. A 2.5 inch CD is available for
laptops, but solid state makes no noise, and has no indicator light. A
different SATA data cable that used to run a failed CD, made no difference.

[William Unruh]

Post by Doug Laidlaw

Post by Bit Twister
If it were me I would get/burn/load the latest systemrescue-8.07-amd64.iso
Click gparted in the bottom tray icom and see if gparted can create a disk
partition table, a partition and format it.

Did that. I have a recent copy of system rescue. Also tried the one for
Mageia. It can't find a drive either. I was thinking that once I have
a partition table, I would be set, but I can't get that far, even.
I still have Rescatux, but it seems to be pretty useless. Years ago, I
had a rescue disk with all the manufacturer's setup tools, but no
longer. Checking the cables next. A 2.5 inch CD is available for
laptops, but solid state makes no noise, and has no indicator light. A
different SATA data cable that used to run a failed CD, made no difference.

Not sure what the sata cable is for for a solid state disk. They usually
mount on the motherboard. Is this an external drive coming in through a
usb port?

The other thing is, as suggested, you might just have a dud drive.
If the system does not recognize the disk, not even any entry under
/dev/sd... or /dev/nvme.... then you have either installed it
incorrectly or you have a dead drive.

[red floyd]

Post by William Unruh
Not sure what the sata cable is for for a solid state disk. They usually
mount on the motherboard. Is this an external drive coming in through a
usb port?

There are SSDs in a 2.5 or 3.5 inch form factor with a SATA connector, as
well as NVME drives, which connect directly to the motherboard.

[TJ]

Post by William Unruh

Post by Doug Laidlaw
I still have Rescatux, but it seems to be pretty useless. Years ago, I
had a rescue disk with all the manufacturer's setup tools, but no
longer. Checking the cables next. A 2.5 inch CD is available for
laptops, but solid state makes no noise, and has no indicator light. A
different SATA data cable that used to run a failed CD, made no difference.

Not sure what the sata cable is for for a solid state disk. They usually
mount on the motherboard. Is this an external drive coming in through a
usb port?

There are external drives that connect via an eSATA port that is present
on some systems. I have one on the back of one of my desktops. It
connects to the motherboard SATA system for data, but needs external
power for the device. The connection transfers data at SATA speeds,
faster than usb 2.0, more comparable to usb 3.

Post by William Unruh
The other thing is, as suggested, you might just have a dud drive.
If the system does not recognize the disk, not even any entry under
/dev/sd... or /dev/nvme.... then you have either installed it
incorrectly or you have a dead drive.

On my system, any device connected to the eSATA port is mounted as a
"removable" device, just the same as it would be if connected via usb.

TJ

[Bit Twister]

Post by Doug Laidlaw

Post by Bit Twister
If it were me I would get/burn/load the latest systemrescue-8.07-amd64.iso
Click gparted in the bottom tray icom and see if gparted can create a disk
partition table, a partition and format it.

Did that. I have a recent copy of system rescue. Also tried the one for
Mageia. It can't find a drive either. I was thinking that once I have
a partition table, I would be set, but I can't get that far, even.
I still have Rescatux, but it seems to be pretty useless. Years ago, I
had a rescue disk with all the manufacturer's setup tools, but no
longer. Checking the cables next. A 2.5 inch CD is available for
laptops, but solid state makes no noise, and has no indicator light. A
different SATA data cable that used to run a failed CD, made no difference.

I probably missed it but what is the model and vendor name of the device?
I want to look at the user manual.

as root check for it with
hwinfo --short > hwinfo.short
and hwinfo > hwinfo.long

and look through the hwinfo,short and hwinfo,long files.

[Doug Laidlaw]

Post by Bit Twister
I probably missed it but what is the model and vendor name of the device?
I want to look at the user manual.
as root check for it with
hwinfo --short > hwinfo.short
and hwinfo > hwinfo.long
and look through the hwinfo,short and hwinfo,long files.

It is a brand new Western Digital "WD Blue" series, using SATA
connections. The current series seems to be "WD Purple." While it was
connected, the Journal showed SATA errors, probably because it could not
be recognized. These stopped as soon as I unlinked the drive.

"Jan 13 11:39:45 dougshost.douglaidlaw.net kernel: ata4: SATA link down
(SStatus 1 SControl 300)"

(A guy on a forum with the same issue was told to swap his cables around
-- on a laptop! I had already tried that.)

As I mentioned, a YouTube tutorial uses the same drive; only the
capacity is different. The setup screen there is taken straight from
the Western Digital Web site. The drive should be listed in Windows as
an "uninitialised device," not just plain "not there." I haven't tried
updating my BIOS; the last time I tried, mine was treated as obsolete.
gparted can't see it nor can two partition managers for Windows. I am
thinking that I would be better off replacing the outboard drive, or it
may not need replacing: it can be read, but the data is corrupted; no
inode numbers can be found to boot from.

[Bit Twister]

Post by Doug Laidlaw

Post by Bit Twister
I probably missed it but what is the model and vendor name of the device?
I want to look at the user manual.
as root check for it with
hwinfo --short > hwinfo.short
and hwinfo > hwinfo.long
and look through the hwinfo,short and hwinfo,long files.

It is a brand new Western Digital "WD Blue" series, using SATA
connections. The current series seems to be "WD Purple." While it was
connected, the Journal showed SATA errors, probably because it could not
be recognized. These stopped as soon as I unlinked the drive.
"Jan 13 11:39:45 dougshost.douglaidlaw.net kernel: ata4: SATA link down
(SStatus 1 SControl 300)"
(A guy on a forum with the same issue was told to swap his cables around
-- on a laptop! I had already tried that.)

Quick google search on SStatus 1 SControl 300
says to swap cables to working drive to prove not a cable/wiring problem
then there is this link.

https://forum.armbian.com/topic/16030-ata1-sata-link-down-sstatus-0-scontrol-300/

[Doug Laidlaw]

Post by Bit Twister
Quick google search on SStatus 1 SControl 300
says to swap cables to working drive to prove not a cable/wiring problem
then there is this link.
https://forum.armbian.com/topic/16030-ata1-sata-link-down-sstatus-0-scontrol-300/

That was one of the first things I tried.

I can now see the Seagate drive in Windows. To keep Win 10 working, I
have installed a maintenance tool from IoBIT called "Advanced System
Care." It has a full quiver of apps. It called a program called Disk
Doctor,which blanked the disk and handed back an empty drive. Linux
still can't see it, but I can now get it working again. The backup data
will change with time, and I can download the contents of the other two
partitions again. But I am not counting my chickens, just yet.

Thanks everybody for your help.

[TJ]

Post by Doug Laidlaw

Post by Bit Twister
Quick google search on SStatus 1 SControl 300
says to swap cables to working drive to prove not a cable/wiring problem
then there is this link.
https://forum.armbian.com/topic/16030-ata1-sata-link-down-sstatus-0-scontrol-300/

That was one of the first things I tried.
I can now see the Seagate drive in Windows.  To keep Win 10 working, I
have installed a maintenance tool from IoBIT called "Advanced System
Care."  It has a full quiver of apps.  It called a program called Disk
Doctor,which blanked the disk and handed back an empty drive. Linux
still can't see it, but I can now get it working again.  The backup data
will change with time, and I can download the contents of the other two
partitions again. But I am not counting my chickens, just yet.
Thanks everybody for your help.

I've seen the time, just recently in fact, where a drive wasn't
recognized by Mageia if hot-plugged while the system was running, but
was if the system was booted with the drive connected.

This particular drive was an ssd that had been the main drive in a
laptop, but had been replaced with a bigger drive. If I hot-plugged it
into my SATA drive bay Mageia couldn't see it, and neither could gparted
if run from within Mageia. But, if it was in place during the boot,
either Mageia or free-standing gparted could see it.

You might want to try that...

TJ

[Doug Laidlaw]

Post by TJ
I've seen the time, just recently in fact, where a drive wasn't
recognized by Mageia if hot-plugged while the system was running, but
was if the system was booted with the drive connected.

I have seen the opposite. Xfce has a setting, "mount a drive which is
hotplugged." It seems to do exactly that: USB sticks which are present
on bootup are not detected; to have them detected, unplug them and then
put them back. Trying a different USB port is recommended, but usually,
it is unnecessary.