failed to read the system array: -2 / open_ctree failed

Hey.

I have a (physically not accessible) server that runs Debian unstable
with a 5.10 kernel and two SATA HDDs which have a RAID1 btrfs.

Every now and then the server ... well I don't know, but it probably
crashes and then hangs at boot when it tries to mount the root fs
(which is the aforementioned btrfs RAID) with the error shown in the
screenshot.


When I boot from a rescue system (kernel 5.13.1, but only btrfs-progs
v4.20.1) and do a normal and lowmem fsck, nothing is found:
root@rescue ~ # btrfs check /dev/sda2  ; echo $?
Opening filesystem to check...
Checking filesystem on /dev/sda2
UUID: 67e35b5c-3dfd-4b00-909d-88308e6b8d85
[1/7] checking root items
[2/7] checking extents
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
found 42220433408 bytes used, no error found
total csum bytes: 38246716
total tree bytes: 389464064
total fs tree bytes: 213270528
total extent tree bytes: 110657536
btree space waste bytes: 91413374
file data blocks allocated: 2509346316288
 referenced 27977461760
0
root@rescue ~ # btrfs check --mode=lowmem /dev/sda2  ; echo $?
Opening filesystem to check...
Checking filesystem on /dev/sda2
UUID: 67e35b5c-3dfd-4b00-909d-88308e6b8d85
[1/7] checking root items
[2/7] checking extents
[3/7] checking free space cache
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs done with fs roots in lowmem mode, skipping
[7/7] checking quota groups skipped (not enabled on this FS)
found 42220433408 bytes used, no error found
total csum bytes: 38246716
total tree bytes: 389464064
total fs tree bytes: 213270528
total extent tree bytes: 110657536
btree space waste bytes: 91413374
file data blocks allocated: 2509346316288
 referenced 27977461760
0



And the fs mounts just fine.


What I also found in (the rescue system’s) dmesg is:
[Thu Aug 19 00:30:56 2021] ahci 0000:00:11.0: version 3.0
[Thu Aug 19 00:30:56 2021] ahci 0000:00:11.0: AHCI 0001.0100 32 slots 6 ports 3 Gbps 0x3f impl SATA mode
[Thu Aug 19 00:30:56 2021] ahci 0000:00:11.0: flags: 64bit ncq sntf ilck pm led clo pmp pio slum part ccc 
[Thu Aug 19 00:30:56 2021] scsi host2: ahci
[Thu Aug 19 00:30:56 2021] scsi host3: ahci
[Thu Aug 19 00:30:56 2021] scsi host4: ahci
[Thu Aug 19 00:30:56 2021] scsi host5: ahci
[Thu Aug 19 00:30:56 2021] scsi host6: ahci
[Thu Aug 19 00:30:56 2021] scsi host7: ahci
[Thu Aug 19 00:30:56 2021] ata3: SATA max UDMA/133 abar m1024@0xfe8ffc00 port 0xfe8ffd00 irq 22
[Thu Aug 19 00:30:56 2021] ata4: SATA max UDMA/133 abar m1024@0xfe8ffc00 port 0xfe8ffd80 irq 22
[Thu Aug 19 00:30:56 2021] ata5: SATA max UDMA/133 abar m1024@0xfe8ffc00 port 0xfe8ffe00 irq 22
[Thu Aug 19 00:30:56 2021] ata6: SATA max UDMA/133 abar m1024@0xfe8ffc00 port 0xfe8ffe80 irq 22
[Thu Aug 19 00:30:56 2021] ata7: SATA max UDMA/133 abar m1024@0xfe8ffc00 port 0xfe8fff00 irq 22
[Thu Aug 19 00:30:56 2021] ata8: SATA max UDMA/133 abar m1024@0xfe8ffc00 port 0xfe8fff80 irq 22
[Thu Aug 19 00:30:56 2021] ata8: SATA link down (SStatus 0 SControl 300)
[Thu Aug 19 00:30:56 2021] ata7: SATA link down (SStatus 0 SControl 300)
[Thu Aug 19 00:30:56 2021] ata6: SATA link down (SStatus 0 SControl 300)
[Thu Aug 19 00:30:56 2021] ata5: SATA link down (SStatus 0 SControl 300)
[Thu Aug 19 00:30:56 2021] ata4: softreset failed (device not ready)
[Thu Aug 19 00:30:56 2021] ata4: applying PMP SRST workaround and retrying
[Thu Aug 19 00:30:56 2021] ata3: softreset failed (device not ready)
[Thu Aug 19 00:30:56 2021] ata3: applying PMP SRST workaround and retrying

=> the "softreset failed (device not ready)"

with ata3 and ata4 being the disks:
[Thu Aug 19 00:30:57 2021] ata4: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[Thu Aug 19 00:30:57 2021] ata4.00: ATA-8: TOSHIBA DT01ACA100, MS2OA750, max UDMA/133
[Thu Aug 19 00:30:57 2021] ata4.00: 1953525168 sectors, multi 0: LBA48 NCQ (depth 32), AA
[Thu Aug 19 00:30:57 2021] ata4.00: configured for UDMA/133
[Thu Aug 19 00:30:57 2021] powernow_k8: Found 1 AMD Athlon(tm) 64 X2 Dual Core Processor 6000+ (2 cpu cores) (version 2.20.00)
[Thu Aug 19 00:30:57 2021] ata3: SATA link up 3.0 Gbps (SStatus 123 SControl 300)
[Thu Aug 19 00:30:57 2021] ata3.00: ATA-8: TOSHIBA DT01ACA100, MS2OA750, max UDMA/133
[Thu Aug 19 00:30:57 2021] ata3.00: 1953525168 sectors, multi 0: LBA48 NCQ (depth 32), AA
[Thu Aug 19 00:30:57 2021] ata3.00: configured for UDMA/133

Could that be the problem? I mean that there is some timing issue, the
kernel tries to bring up the devices but that fails at first, while it
succeeds later (but then mounting the root fs already failed)?



Maybe it's also some hardware issue (though the provider has already
replaced cables and a memtest didn't reveal anything useful).


Any ideas what the errors (failed to read the system array: -2 /
open_ctree failed) could indicate?

Thanks,
Chris.