(Solved) [6.4.0] No network overnight. Edit: lockups

[michael123]

Hi

 

I upgraded tonight to 6.4.0, played with it for half an hour yesterday and it was fine.

 

This morning I couldn't connect, neither to SMB nor via web page

 

Did hardware reboot

 

What could be wrong?

[michael123]

Continued:

 

After reboot and mounting the array, started parity check. Noticed that ~180 MB/sec speed for 6.3.5 dropped to below 70 MB/sec in 6.4.0

But the system after 5 minutes locked up. Didn't see any lights of the working disks during parity process, and of course no network

 

Some kind of lock-up

 

Went back to 6.3.5 and re-run the parity check

 

So far so good

 

 

[bonienl]

diagnostics are required for more information. These should be obtained before restarting.

 

Do you use a Ryzen processor? If so look at the specific topics and workarounds.

 

[michael123]

It is a lock-up, there isn't much I could do before hitting on the power button

 

I have Supermicro X11SSH-F with Xeon  E3-1270 V6 and 64GB Samsung ECC memory.

 

Strange that when I connected to IPMI port, the console image was not shown as well (?)

 

 

 

 

[JorgeB]

1 minute ago, michael123 said:

Strange that when I connected to IPMI port, the console image was not shown as well (?)

 

Known Issues

• Certain motherboards with on-board Aspeed IPMI graphics adaptors may lose video or switch to different colored text during the boot process. To prevent this, add the nomodeset kernel option to your syslinux append line.

[michael123]

Just now, johnnie.black said:

 

Known Issues

• Certain motherboards with on-board Aspeed IPMI graphics adaptors may lose video or switch to different colored text during the boot process. To prevent this, add the nomodeset kernel option to your syslinux append line.

 

Thanks, that will resolve the console issue on IPMI, but what's for lockups? Am I the only one here?

[bonienl]

Is it a lockup and both GUI and SSH are not working?

 

Did you try to start in safemode?

 

[michael123]

Just now, bonienl said:

Is it a lockup and both GUI and SSH are not working?

 

Did you try to start in safemode?

 

 

Not just GUI and SSH did not work, the parity process, which started after mounting due to the hard reboot, stuck after 5 min (!)

 

I did not try safemode

 

Besides the lockup, there is an obvious speed issue, could it be related to the recent cyber processor fixes?

Do I need them at all?

 

 

 

[JorgeB]

Try safemode

[michael123]

The only thing there I have is CrashPlan docker & NerdTools

Everything else is stock, can you elaborate what should I do/see then?

 

It is a "production" system with all my data on it (albeit backed up in the cloud for most critical files), the parity halt raises an eyebrow.

 

 

 

[bonienl]

23 minutes ago, michael123 said:

Besides the lockup, there is an obvious speed issue

 

The usual cause of a speed issue is a bad or poor performing hard disk.

 

When starting in safemode we can exclude any plugin from being the culprit. E.g. NerdTools may install several (incompatible?) packages.

 

Ps. My last parity check on 6.4.0 did not show a decrease in speed.

 

[michael123]

I have a 'screen' from NerdTools, unlikely it will cause any issue (I used it merely when I did rsync to upgrade from reiser to xfs)

 

Parity speed as well as stability is now back to normal since I downgraded to 6.3.5

 

Due to lock-up, to me it seems like something related to this Spectre fix

 

[michael123]

I removed the S3 Sleep plugin (don't use it anyway since I have now crashplan), and fingers crossed system is now operational for about 4 hours running parity,.