On 11/2/23 01:05, gene heskett wrote:
On 11/2/23 01:46, David Christensen wrote:
On 11/1/23 18:34, gene heskett wrote:
So what i'm going to do next is transfer my /home, the whole MaryAnn
from a 4 drive raid10 to a single 2T SSD,, and then switch /home from
the raid to a single drive, thereby removing the raid10 from the
culprit list.
The df(1) output, below, indicates that /home is still on RAID10.
If that doesn't fix it, then its something to do with the graficks.
I have one real clue, some one the openscad list said the preview
mode opened a bunch of gfx buffers it used the preview mode that it
didn't do in final build mode, so I ran some tests on a fairly
complex project, noting that the openscad logging window said a
preview was done in a bit over a second but the final monocolored
version took over 9 seconds to render, giving me an image I could
move in 3d space in the preview window in real time. But that 1.2
second preview render took 30 to 45 seconds to show in\on screen and
every update of that image took at least 45 seconds to register on
screen, and during that 45 seconds or more, openscad was frozen from
all other input. That is as close to a clue as I have, also
manifested when any kind of a file requester needs to be drawn on
screen it is subjectto to that same time killing lockup except it
might be as long as 5 minutes before the requestor is drawn on
screen. That raid was built by buster, worded fine for bullseye, but
seems to be a disaster for bookworm, yet an fsck finds nothing wrong.
Those are the clues I have. Nothing obvious and common to all this
contaminates the logs.
I own the whole nearly 1.8T raid10, every byte of it. Why can't I
use it? Instead its holding me for an undefined ransom in time that I
at 89 don't have a lot of left.
I assume this is the computer with the Asus PRIME Z370-A II
motherboard, Intel i5-9600K processor, 32 GB RAM, WD Black 2 TB NVMe
PCIe 3.0 x4 SSD, a few PCIe x1 SATA III HBA's with many ports, many
SATA III SSD's and/or HDD's, and a few md RAID (?).
Yup, I just added another 3.6Tb of sdd's that are still unformatted.
Gigastone's from Taiwan.
4 @ 1 TB or 2 @ 2 TB? If and when you configure them, please note that
in future posts.
Did you connect the /home RAID10 4 @ 1 TB SSD's to one PCIe x1 HBA?
I assume its an X1 card, small footprint, cheap version with 10 bondouts
but only 6 have drive cable sockets installed. Only 4 for the raid10 are
actually used.
One PCIe 3.0 lane is rated for 985 MB/s, which is less than the sum of
the throughputs for 4 @ SATA III SSD's. But, it should not cause the
openscad issues.
I have another card I haven't tried, is a full 16 ports
but is also the narrow pci-e footprint. Neither is sold as a raid card,
so its all mdadm.
I recall discussing that about two months ago. Similar to the 6 port
HBA, it will limit the aggregate throughput; but should not cause
openscad issues.
This worked perfectly for buster and bullseye, so it seems this problem
is bookworms problem.
If you have a Buster or Bullseye drives or system images, they could be
useful for testing.
Are your file systems full?
/home is currently about 20% used, everything else is much lower yet.
# df /boot / /tmp /home
The whole thing:
gene@coyote:~$ df
Filesystem 1K-blocks Used Available Use% Mounted on
udev 16329748 0 16329748 0% /dev
tmpfs 3272676 1880 3270796 1% /run
/dev/sda1 863983352 15989096 804032608 2% /
tmpfs 16363376 1232 16362144 1% /dev/shm
tmpfs 5120 8 5112 1% /run/lock
/dev/sda3 47749868 260 45291600 1% /tmp
/dev/md0p1 1796382580 327360472 1377697132 20% /home
tmpfs 3272672 3768 3268904 1% /run/user/1000
Okay -- no free space issues.
What drive corresponds to /dev/sda? If not the WD Black 2 TB NVMe
drive, then is the WD Black installed and configured?
Do your file systems need defragmenting?
# e4defrag -c /boot / /home
dnk, wasn't fam with how to do it, And worst case score for that cmd is 1.
/home was the last check:
<Fragmented files> now/best size/ext
1.
/home/gene/.mozilla/firefox/f8j7d2lj.default-esr/storage/default/https+++www.wsaz.com/ls/data.sqlite
7/1 4 KB
2.
/home/gene/.mozilla/firefox/f8j7d2lj.default-esr/storage/default/https+++apps.sascdn.com/ls/data.sqlite
7/1 4 KB
3.
/home/gene/.mozilla/firefox/f8j7d2lj.default-esr/storage/default/https+++www.euronews.com/ls/data.sqlite
6/1 4 KB
4.
/home/gene/.mozilla/firefox/f8j7d2lj.default-esr/storage/default/https+++www.britannica.com/ls/data.sqlite
6/1 4 KB
5.
/home/gene/.mozilla/firefox/f8j7d2lj.default-esr/storage/default/https+++travelermaster.com/ls/data.sqlite
6/1 4 KB
Total/best extents 1245442/1166289
Average size per extent 262 KB
Fragmentation score 1
[0-30 no problem: 31-55 a little bit fragmented: 56- needs defrag]
This directory (/home) does not need defragmentation.
Done.
Okay -- no fragmentation issues.
Thank you David.
YW. :-)
The "PRIME Z370-A II Series" user's manual indicates various RAID and
non-RAID storage configurations using PCIe slots, PCIe Hyper M.2 cards,
M.2 sockets, M.2 drives, SATA 6.0 Gb/s ports, SATA drives, and Setup
NVRAM settings. The details are non-trivial -- perhaps your system is
mis-configured (?):
https://www.asus.com/us/supportonly/prime%20z370-a%20ii/helpdesk_manual/
Please examine the following NVRAM settings:
BIOS Setup
Advanced menu
PCH Storage Configuration
SATA Mode Selection AHCI
If the WD Black is installed in socket M.2_1:
BIOS Setup
Advanced menu
Onboard Devices Configuration
M.2_1 Configuration PCIE mode
If the WD Black is installed in socket M.2_2 (ports SATA6G_5 and
SATA6G_6 will be disabled in X4 mode):
BIOS Setup
Advanced menu
Onboard Devices Configuration
M.2_2 PCIe Bandwidth Configuration X4
To check the hardware and NVRAM configuration, tear down the system to
just the motherboard and one SATA drive with Bookworm, reset Setup NVRAM
to defaults, and test. If the openscad issue is present, try Bullseye
and/or Buster. If the issue is absent, continue adding hardware,
configuring NVRAM, and testing until either you find the problem or the
system is complete.
David
