Re: strange behaviour after upgrading from Jessie to Stretch
On 14/07/17 11:44 PM, Gary Dale wrote:
My "server" is a bit of an antique. It's a Gigabyte MA790GP-UD4H with
a Phenom II 940 processor, an SSD and 4 2T drives in a RAID 6 array
plus 8G of DDR2 ram.
A while back I found it necessary to add a Vantec PCIe SATA 6Gbps card
with 2 internal SATA ports after ports 4 & 5 on the motherboard
stopped recognizing the drives attached to them. It's been running
with that configuration for at least a year.
This morning I upgraded from Jessie to Stretch and things went
haywire. I could successfully boot into the system using
SystemRescueCD but the 2 drives attached to the add-in card weren't
recognized. My RAID 6 array was running degraded on the two drives
attached to ports 2 & 3. Attempts to boot into Debian/Stretch were
largely unsuccessful, with the computer usually locking up after
reporting an error on an ATA device.
After a lot of fiddling I discovered that the HDs themselves were all
OK but could only be accessed using SATA ports 0 - 3. However my
boot/system SSD and the BluRay writer both operate fine off of ports
4 & 5. They also work when attached to the add-in card. I removed the
add-in card and it now successfully boots into Debian/Stretch.
Both SystemRescueCD and Debian/Stretch are using 4.9 kernels but the
former seemed able to boot while the latter wouldn't. I'm also
confused by the problems I've been having with HDs not working when
attached to some ports. Can any hardware gurus explain what's going on?
On Saturday my "server" started beeping at me - intermittent long beeps
with no discernible pattern. I eventually replaced the power supply and
that seems to have stopped the beeping. However I can no longer boot
into Debian/Stretch (4.9.0 kernel). I can only boot into the older
Jessie 3.16 kernel.
Moreover the RAID 6 array was running degraded. Trying out various
permutations of drives using the 6 onboard SATA ports didn't help. I had
to put the Vantec card back in and plug 2 HDs into it.
Now that the RAID 6 arrays have rebuilt, I tried booting into the 4.9.0
kernel again. It didn't go well. I left the machine for a couple of
hours and it still hadn't booted. It wasn't frozen but it also wasn't
working. There were some messages on the screen about jiffies and other
things - not the usual boot messages. I could hit enter and the screen
would scroll up so things weren't locked up.
I booted back into 3.16 and things started normally - back to the login
prompt in a matter of seconds after selecting the older kernel.
There is nothing in syslog covering the period that the system was
booting into 4.9 so it didn't get that far.
I turned on persistent storage of log messages but I get nothing after
the shutdown of the previous 3.16 session and the start of the next
3.16. The boot attempt into 4.9 isn't there. journalctl --list-boots shows:
-1 d7310cc3b62c4586a698c8df41471bde Tue 2017-07-25 15:44:21 EDT—Tue
2017-07-25 18:41:35 EDT
0 5b60c56ab0674f98ae85bd900616bc38 Tue 2017-07-25 18:44:23 EDT—Tue
2017-07-25 18:45:37 EDT
Any ideas?
Reply to: