You now have a slow access to one/more of your RAID devices.
Hi Gene,Correct, a one size fits all machine.
Frankly: Dealing with you over a mailing list can be very frustrating for others trying to help (and especially for people trying to follow the list who are reading the lists in the background and facing long, long threads).
You're not helping explain yourself well because the mails keep referring
to "other stuff that happened a while ago".
People ask to see mails / messages or whatever exactly because we can't
sit next to you, we can't see what you type, we can't see what you're meaning.
That's why well-meaning folk keep asking questions to try
and establish what's going on and get a sense of where we can help
(if at all).
There's a whole series of long threads which loop through several
subjects - I can tease out a couple of things.
1.) You have one large deskside machine - large enough that it's tough
to lift or move - which is used for many things.
2.) You've added various drive controllers and various drives over a
period.
Unclear: At least one of your RAID devices may be mixed between on motherboard connections and on-card drive controller connections??No, I boot from /dev/sda, a 1T samsung 870 SSD plugged into the
3.) You "lost" a RAID a while ago so you don't trust RAID on some devicesNo here, it was a pair of quite new 2T seagates that died and started
but you're persisting with RAIDs.
4.) You have various add in cards but you don't seem to know which RAID is which / what's "locking" your filesystem / what's causing your problems.Which from the very limited clues seems to be related to my original of
You now have a slow access to one/more of your RAID devices.
5). Unclear: All / ("most"??) of those RAID devices are using Linux mdadm rather than "RAID" supplied by the individual cards/controllers.
Various of us - including myself - have suggested that you simplify things
/ get another machine and divide up functionality. For various reasons
you can't / won't do that.
Can you answer the questions I've posted above, please, to tryInstant /etc/fstab:
and clarify what you have. I would have asked you for /etc/fstab and
a couple of other files, but this is good enough to be going on with.
All the best, as ever,Likewise Andy, take care, stay warm, dry and well.
Andy Cater
(amacater@debian.org)
.
# first put it where it is now & reboot...
#LABEL=homesde1 /mnt/homesde1 ext4 errors=remount-ro 0 2
I have not been able to use that last line as a target for rsync
On 1/14/24 06:59, Andrew M.A. Cater wrote:
Hi Gene,
There's a whole series of long threads which loop through several
subjects - I can tease out a couple of things.
1.) You have one large deskside machine - large enough that it's toughCorrect, a one size fits all machine.
to lift or move - which is used for many things.
2.) You've added various drive controllers and various drives over a period.
Correct.
Unclear: At least one of your RAID devices may be mixed between on motherboard connections and on-card drive controller connections??No, I boot from /dev/sda, a 1T samsung 870 SSD plugged into the motherboards sata which has 6 ports
Because I didn't have 4 ports left for the raid, it on its own controller, one of 2 extra sata controllers currently plugged in. Both of the extra controlleres are just controllers, no raid in their pedigree, 1st extra has
6 ports, 2nd extra has 16 ports.
3.) You "lost" a RAID a while ago so you don't trust RAID on some devices but you're persisting with RAIDs.No here, it was a pair of quite new 2T seagates that died and started this whole maryann. Lasted about a month from 1st powerup to going offline in the night with no warning about 3 weeks after 1st powerup. Lost everything back to about 2002. The only raid I've ever had is the current one, which
smartctl was sending me emails about but not thru a normal chaanel, I only found them when I found a strange mbox file in my home dir. Last mail in the mbox file was dated Jan 7th of this year.
But I've now sussed the smartctl syntax and all 4 drives of the raid say
they are healthy.
4.) You have various add in cards but you don't seem to know which RAID is which / what's "locking" your filesystem / what's causing your problems.
You now have a slow access to one/more of your RAID devices.Which from the very limited clues seems to be related to my original of of plasma for a desktop, with xfce4 on top of that. So I suspecting the problem might be mixed gui related. This lag or lockup, whatever you want to call it occurs for any app that opens a file requestor, there at least 30 seconds of this lag before the gui opens the requestor, at which point everything returns to normal. Failing ns reslution? I've NDI. The lags are not logged anyplace I've managed to find a log to read.
5). Unclear: All / ("most"??) of those RAID devices are using Linux mdadm rather than "RAID" supplied by the individual cards/controllers.
Correct.
Various of us - including myself - have suggested that you simplify things / get another machine and divide up functionality. For various reasons
you can't / won't do that.
Mostly lack of space in this tiny childs bedroom to do that, over the last
35 years its best described as a midden heap. ;o)>
Can you answer the questions I've posted above, please, to tryInstant /etc/fstab:
and clarify what you have. I would have asked you for /etc/fstab and
a couple of other files, but this is good enough to be going on with.
gene@coyote:/etc$ cat fstab
# /etc/fstab: static file system information.
#
# Use 'blkid' to print the universally unique identifier for a
# device; this may be used with UUID= as a more robust way to name devices
# that works even if disks are added and removed. See fstab(5).
#
# systemd generates mount units based on this file, see systemd.mount(5).
# Please run 'systemctl daemon-reload' after making changes here.
#
# <file system> <mount point> <type> <options> <dump> <pass>
# / was on /dev/sda1 during installation UUID=f295334b-fdcb-4428-bed3-cb9e9e129be6 / ext4 errors=remount-ro 0 1
# /tmp was on /dev/sda3 during installation UUID=518cb65d-21f0-493f-8bb5-a5f435796991 /tmp ext4 defaults
0 2
# swap was on /dev/sda2 during installation UUID=422b50db-9913-4ed3-92c3-dc18be72cc61 none swap sw
0 0
/dev/sr0 /media/cdrom0 udf,iso9660 user,noauto 0 0 UUID=bc6135de-0578-4e3b-b2c0-5c4687abd9bd /home ext4 errors=remount-ro 0 2
UUID=d24c3a99-9f40-4b71-92d4-916804553cb5 none swap sw 0 0 # first put it where it is now & reboot
#LABEL=homesde1 /mnt/homesde1 ext4 errors=remount-ro 0 2
gene@coyote:/etc$
I have not been able to use that last line as a target for rsync, it make around 13.5 gig of a 360G copy and locks up all i/o. So I've now refomatted the drive, an am about to powerdown and move that drives data cable to the 2nd, 16 port controller. When plugged into the 1st 6 port controller it ran to completion but not to the drive mounted by its label. So that made me
move that data cable off the 1st controller to an empty plug on the motherboard And that crashes the system i/o. mouse works, no bash shell or keyboard works, so now I'm going to test by connecting that drive to the
2nd, 16 port controller and see how far rsync gets. The current idea is to remove the raid from the environment by copying it to a single drive, and editing fstab to make that copy my /home partition. if that does not remove this lag, then the gui mix theory as the src of this lag is re-enforced. if the lag is gone, then its in the raid. And we are 1 clue closer to finding the src of this problem. And ALL of this caused by the installer putting orca and brltty in without asking just because it found an fdti serial adapter connected to a cm11a x-10 controller for my Christmas lights. Listening to orca pronounce each and every keystroke as you type is enough
to drive one to drink and its a very short drive when 20+ installs have been done because once installed, it cannot be disabled by any means and still reboot, init locks up waiting for orca, so a reboot was a reinstall. And I catch hell for squawking about it at the time when bookworm was only weeks old. I rest my case, I'm powering down to move a data cable and make the
next test.
All the best, as ever,Likewise Andy, take care, stay warm, dry and well.
Andy Cater
(amacater@debian.org)
.
Cheers, Gene Heskett.
--
"There are four boxes to be used in defense of liberty:
soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author, 1940)
If we desire respect for the law, we must first make the law respectable.
- Louis D. Brandeis
gene heskett composed on 2024-01-14 12:04 (UTC-0500):
# first put it where it is now & reboot...
#LABEL=homesde1 /mnt/homesde1 ext4 errors=remount-ro 0 2
I have not been able to use that last line as a target for rsync
That's not unexpected. /mnt/ is intended for /temporary/ or /transient/ mounting,
while /etc/fstab is OTOH intended for routine.
The explosion could have occurred
by inserting a USB stick while rsync was running and you were engaging in root
activities. As regular user, most DEs now use /run/media/<user> instead of /tmp/.
Best anyway to find someplace besides your /mnt/ tree for that filesystem, maybe
/home/coyotebak/ or /backupdisk/.
On Sun, Jan 14, 2024 at 12:33:39PM -0500, Felix Miata wrote:
gene heskett composed on 2024-01-14 12:04 (UTC-0500):
# first put it where it is now & reboot...
#LABEL=homesde1 /mnt/homesde1 ext4 errors=remount-ro 0 2
I have not been able to use that last line as a target for rsync
That's not unexpected. /mnt/ is intended for /temporary/ or /transient/ mounting,
while /etc/fstab is OTOH intended for routine.
How should the mount point have an influence on transfer rates?
The explosion could have occurred
by inserting a USB stick while rsync was running and you were engaging in root
activities. As regular user, most DEs now use /run/media/<user> instead of /tmp/.
Best anyway to find someplace besides your /mnt/ tree for that filesystem, maybe
/home/coyotebak/ or /backupdisk/.
You think an automounter mounted some stuff beneath /mnt/?
I think they don't do that for the last twenty years, at least
(before /run/media/<user> it has been /media/<user> for quite
a while already...
On Sun, Jan 14, 2024 at 11:58:42AM +0000, Andrew M.A. Cater wrote:
You now have a slow access to one/more of your RAID devices.
Does he, though? I thought we had established some time late last
year that his *symptom* (delayed startup of some applications) had
nothing at all to do with his storage devices, and everything to do
with his desktop environment.
gene heskett composed on 2024-01-14 12:04 (UTC-0500):
# first put it where it is now & reboot...
#LABEL=homesde1 /mnt/homesde1 ext4 errors=remount-ro 0 2
I have not been able to use that last line as a target for rsync
That's not unexpected. /mnt/ is intended for /temporary/ or /transient/ mounting,
while /etc/fstab is OTOH intended for routine. The explosion could have occurred
by inserting a USB stick while rsync was running and you were engaging in root
activities. As regular user, most DEs now use /run/media/<user> instead of /tmp/.
Best anyway to find someplace besides your /mnt/ tree for that filesystem, maybe
/home/coyotebak/ or /backupdisk/.
https://en.wikipedia.org/wiki/Filesystem_Hierarchy_StandardOr are you saying I should mkdir that mount point in the rad10, and then
tomas composed on 2024-01-14 19:15 (UTC+0100):
On Sun, Jan 14, 2024 at 12:33:39PM -0500, Felix Miata wrote:
gene heskett composed on 2024-01-14 12:04 (UTC-0500):
# first put it where it is now & reboot...
#LABEL=homesde1 /mnt/homesde1 ext4 errors=remount-ro 0 2
I have not been able to use that last line as a target for rsync
That's not unexpected. /mnt/ is intended for /temporary/ or /transient/ mounting,
while /etc/fstab is OTOH intended for routine.
How should the mount point have an influence on transfer rates?
AFAIK, nothing I wrote would be expected to have any relationship to transfer rates. My point was entirely about suitability of /mnt/ for fstab entries.
The explosion could have occurred
by inserting a USB stick while rsync was running and you were engaging in root
activities. As regular user, most DEs now use /run/media/<user> instead of /tmp/.
Best anyway to find someplace besides your /mnt/ tree for that filesystem, maybe
/home/coyotebak/ or /backupdisk/.
You think an automounter mounted some stuff beneath /mnt/?
I think they don't do that for the last twenty years, at least
(before /run/media/<user> it has been /media/<user> for quite
a while already...
I don't have a working knowledge of all the deviations from FHS or other standards
that Gene employs, and neither am I familiar with behaviors of DEs I do not use.
When one has /mnt/ in fstab, where would one put a transient manual mount? Another
would need to be created, lest done to /mnt/ on coyote, /mnt/homesde1/'s filesystem would disappear, no trivial danger in the context of deteriorated short
term memory.
Felix Miata wrote:...
I have mount
points scattered about this system, literaaly all over that just work,
since when is /mnt some special thing?
...https://en.wikipedia.org/wiki/Filesystem_Hierarchy_StandardOr are you saying I should mkdir that mount point in the rad10, and then mount one of these SSD's to it? Sounds like the long way around the
bush but it might work, I'll try it. But that would be forever recursive
w/o excluding that dir from the copy.
Felix Miata wrote:
AFAIK, nothing I wrote would be expected to have any relationship to transferAnd my point is that for a one time copy, its was handy. I didn't have
rates. My point was entirely about suitability of /mnt/ for fstab entries.
to mkdir a mount point for it.
This whole thing has just one objective, making a copy of the raid10Evolution as taught in public schools is, like religion,
/home onto a single drive that I could us to edit the raid out of fstab, substituting the single drive copy. This raid has 2 of its 4 drives complaining to smartctl. Get my data off it, by whatever means works.--
/home/coyotebak would be in the raid, but something in the system /backupdisk/ as a mount point would not be in the raid. But I have
mount points scattered about this system, literaaly all over that just
work, since when is /mnt some special thing? its just an empty dir I
can mount any block thing to. one furinstance is an /sshnet directory,
inside of which is a few nore directories that I mount the rest of my
cnc and 3d printers to using sshfs, so they are a direct link to the
/home/me directories of every machine on the premises. I have a script
in my private bin directory that mounts them all. I get tired of
repeating my user pw while the script is running, but it just works.
gene heskett composed on 2024-01-14 18:15 (UTC-0500):so I should use /media? In any event that line is now commented out of
Felix Miata wrote:...
I have mount
points scattered about this system, literaaly all over that just work,
Fine! It's your stuff.
since when is /mnt some special thing?
Since 1994, 30 years ago next month:
... http://www.ibiblio.org/pub/Linux/docs/fsstnd/old/fsstnd-1.1/fsstnd-1.1.txt
...https://en.wikipedia.org/wiki/Filesystem_Hierarchy_StandardOr are you saying I should mkdir that mount point in the rad10, and then
mount one of these SSD's to it? Sounds like the long way around the
bush but it might work, I'll try it. But that would be forever recursive
w/o excluding that dir from the copy.
I'm only suggesting you find a place other than /mnt/ for anything found in /etc/fstab, based upon the definition of /mnt/ in FHS. Conforming your machinery
to FHS is not mandatory, just recommended, a good idea.
gene heskett composed on 2024-01-14 18:39 (UTC-0500):I don't recall ever mounting something to /mnt, always to a subdir in mount.
Felix Miata wrote:
AFAIK, nothing I wrote would be expected to have any relationship to transferto mkdir a mount point for it.
rates. My point was entirely about suitability of /mnt/ for fstab entries. >> And my point is that for a one time copy, its was handy. I didn't have
/mnt/ is intended for one-time copies, just the ticket for that particular exercise.
subdirectories in /mnt/ unless a filesystem is temporarily mounted there. It's
allowed, but not particularly a product of wisdom, particularly if you forget to
undo it before rebooting without the configured filesystem available, or an unexpected reboot occurs first. Remember your short-term memory quality?
This whole thing has just one objective, making a copy of the raid10Evolution as taught in public schools is, like religion,
/home onto a single drive that I could us to edit the raid out of fstab,
substituting the single drive copy. This raid has 2 of its 4 drives
complaining to smartctl. Get my data off it, by whatever means works.--
based on faith, not based on science.
Team OS/2 ** Reg. Linux User #211409 ** a11y rocks!
Felix Miata
On Sun 14 Jan 2024 at 18:15:13 (-0500), gene heskett wrote:
/home/coyotebak would be in the raid, but something in the system
/backupdisk/ as a mount point would not be in the raid. But I have
mount points scattered about this system, literaaly all over that just
work, since when is /mnt some special thing? its just an empty dir I
can mount any block thing to. one furinstance is an /sshnet directory,
inside of which is a few nore directories that I mount the rest of my
cnc and 3d printers to using sshfs, so they are a direct link to the
/home/me directories of every machine on the premises. I have a script
in my private bin directory that mounts them all. I get tired of
repeating my user pw while the script is running, but it just works.
When I save a file, I use bash-completion to help write directory
names etc, but I know where I'm going, so to speak. But I have this
idea at the back of my mind that when a GUI dialog box opens up for
you to save a file, it has an extensive look around the filesystem.
Direct links (symlinks, I assume) to networked machines is a great
way of slowing this process down, and any unresolvable names will
make things far worse. Is this a possible cause for your problem?
Cheers,
David.
Felix Miata wrote:
I'm only suggesting you find a place other than /mnt/ for anything found in >> /etc/fstab, based upon the definition of /mnt/ in FHS. Conforming your machinery
to FHS is not mandatory, just recommended, a good idea.
so I should use /media?
tomas composed on 2024-01-14 19:15 (UTC+0100):
I have not been able to use that last line as a target for rsync
That's not unexpected. /mnt/ is intended for /temporary/ or /transient/ mounting,
while /etc/fstab is OTOH intended for routine.
How should the mount point have an influence on transfer rates?
AFAIK, nothing I wrote would be expected to have any relationship to transfer rates. My point was entirely about suitability of /mnt/ for fstab entries.
/home/coyotebak would be in the raid, but something in the system /backupdisk/ as a mount point would not be in the raid. But I have mount points scattered about this system, literaaly all over that just work, since when is /mnt some special thing? [...]
Or are you saying I should mkdir that mount point in the rad10, and then mount one of these SSD's to it? [...]
On 1/14/24 18:57, Felix Miata wrote:
gene heskett composed on 2024-01-14 18:39 (UTC-0500):
Felix Miata wrote:
My point was entirely about suitability of /mnt/ for fstab entries.And my point is that for a one time copy, its was handy. I didn't have
to mkdir a mount point for it.
/mnt/ is intended for one-time copies, just the ticket for that particular exercise.I don't recall ever mounting something to /mnt, always to a subdir in mount.
On Sun 14 Jan 2024 at 20:15:16 (-0500), gene heskett wrote:
On 1/14/24 18:57, Felix Miata wrote:
gene heskett composed on 2024-01-14 18:39 (UTC-0500):I don't recall ever mounting something to /mnt, always to a subdir in mount.
Felix Miata wrote:
My point was entirely about suitability of /mnt/ for fstab entries.And my point is that for a one time copy, its was handy. I didn't have >>>> to mkdir a mount point for it.
/mnt/ is intended for one-time copies, just the ticket for that particular >>> exercise.
How did you mount to a subdir in /mnt without making a mount point?
Your two statements appear contradictory.
Cheers,
David.
.
On 1/15/24 14:57, David Wright wrote:
On Sun 14 Jan 2024 at 20:15:16 (-0500), gene heskett wrote:
On 1/14/24 18:57, Felix Miata wrote:
gene heskett composed on 2024-01-14 18:39 (UTC-0500):
Felix Miata wrote:
My point was entirely about suitability of /mnt/ for fstab entries.And my point is that for a one time copy, its was handy. I didn't have
to mkdir a mount point for it.
/mnt/ is intended for one-time copies, just the ticket for that particularI don't recall ever mounting something to /mnt, always to a subdir in mount.
exercise.
How did you mount to a subdir in /mnt without making a mount point?
Your two statements appear contradictory.
On this particular ball of rock and water at least, called a planet in
most circles, you "mkdir /mnt/whatever". You don't have to mount
directly on /mnt and I don't think I ever have. Do it 50 times with
your own version of "whatever", same with any path to anyplace on the
system. Nothing special about it.
Sysop: | Keyop |
---|---|
Location: | Huddersfield, West Yorkshire, UK |
Users: | 307 |
Nodes: | 16 (2 / 14) |
Uptime: | 106:34:25 |
Calls: | 6,852 |
Calls today: | 3 |
Files: | 12,355 |
Messages: | 5,415,833 |