Can't rebuild parity

[cbr600ds2]

Hello -  

 

I could really use some help.   i don't want to have to buy a second server just to see what's wrong but I'm having some issues w/ this server that has been running pretty good for a while.  I installed a RES2SV240 a week or so ago because I was running 2 9207-8i's and I wanted to go to something else.  Well for some reason it threw a red ball on one of the disks and then I couldn't get the GUI to come up (it kept giving me error 504 after refusing to load the full GUI).  For S&G's I went back to both 9207's to see if didn't like the new card for some reason and every time I would try to rebuild it would do it for a while and then stop and not load.   The disks  were getting hot so I shut it down over night and let it cool down.   Now I've got it going again and I've lost HD temp readings and its like it not loading the full GUI.   I'm wondering if you think the cards/cables are going bad or maybe the MB( I hope not).   I've attached the syslog but I couldn't even get it to download diagnostics.  It just keeps running.  The diagnostics zip is pulled from yesterday.  The syslog  was pulled just now.  It's trying to rebuild but its been really slow.  Now I'm getting the Error 504 which I was getting earlier as well.    

 

Any thoughts?  

 

 

Finally got a new HD - precleared and it was being rebuilt and now its doing the same thing.  No errors on preclear.   Thoughts?  everything was running fine emulated without the disk.   

skynet-syslog-20200814-0100.zip skynet-diagnostics-20200812-1933.zip

Edited August 30, 2020 by cbr600ds2
new HD did not fix.

[trurl]

 

[cbr600ds2]

Well I'll be a monkey's uncle.  I had disabled it (I'm pretty damn sure) when I first installed the Ryzen and didn't do any updating to the bios so I guess it never occurred to me that it turned back on to be honest.  I'm trying it again.  Will update.  

[cbr600ds2]

hi @trurl

 

Didn't work  

 

Disabled the C-states, changed the power supply idle control and did the rcs no call back.  Still locks up.   

 

[trurl]

Have you done memtest?

[JorgeB]

2 hours ago, cbr600ds2 said:

Still locks up. 

Did you also stop overclocking the RAM? On the diags posted it's above max supported speeds, also explained on the link above.

[cbr600ds2]

Going to do a memtest tonight.  I never chose ram overclocking and it was in auto but I'll check that tonight as well.  

[cbr600ds2]

Hello - 

 

RAM is set to auto just as I thought so.  I couldn't do the memtest86+ from the boot screen.  it just kept kicking back to the to reboot loop.  so it's running right now.  Will update.   thanks!  Looking into the RAM thing.  

 

Ok but how do I read this? I have DDR4-3200.  32GB ripjaws in 2x16 config.     AHHH I had it I put the sticks in dual channel mode.  Should I put it back in to single channel mode?  Would that make a difference?  

Edited August 15, 2020 by cbr600ds2

[cbr600ds2]

I just realized that the memory I have isn't "supported" by the mb ram support list.  I wonder if it just realized it now...that may be it, right?

[cbr600ds2]

Aug 13 20:21:21 Skynet kernel: rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
Aug 13 20:21:21 Skynet kernel: rcu:     9-...0: (2 ticks this GP) idle=14a/0/0x1 softirq=75806/75806 fqs=14777 
Aug 13 20:21:21 Skynet kernel: rcu:     (detected by 8, t=60002 jiffies, g=269305, q=19498)
Aug 13 20:21:21 Skynet kernel: Sending NMI from CPU 8 to CPUs 9:
Aug 13 20:21:21 Skynet kernel: NMI backtrace for cpu 9

 

does this mean the chip went bad?

[Vr2Io]

6 hours ago, cbr600ds2 said:

I couldn't do the memtest86+ from the boot screen.  it just kept kicking back to the to reboot loop.

This is know issue, pls setting BIOS boot in Legacy mode ( non-UEFI ) for memory test.

 

2 hours ago, cbr600ds2 said:

mb ram support list

Don't be silly, it is not necessary.

 

Pls note you already in single channel mode, you should reinsert the module for dual channel mode then perform memory test.

Edited August 15, 2020 by Benson

[JorgeB]

7 hours ago, cbr600ds2 said:

Ok but how do I read this?

You have both dimms on the same channel, so it's overloading that channel and running out of spec, it's also running slower than it would in dual channel.

You need to to alternate the slots when installing just two dimms, usually they are color codded so both dimms should be on the same color slots.

[cbr600ds2]

7 hours ago, johnnie.black said:

You have both dimms on the same channel, so it's overloading that channel and running out of spec, it's also running slower than it would in dual channel.

You need to to alternate the slots when installing just two dimms, usually they are color codded so both dimms should be on the same color slots.

I did swap it out and its still hanging.  Memtest ran for 12 hours and indicated no issues.  I'm really curious if its the CPU that took a dump because I keep getting this over and over - 

Aug 15 09:55:27 Skynet kernel: rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
Aug 15 09:55:27 Skynet kernel: rcu:     3-...0: (32 ticks this GP) idle=79a/0/0x1 softirq=81265/81265 fqs=285001 
Aug 15 09:55:27 Skynet kernel: rcu:     (detected by 10, t=1140034 jiffies, g=471857, q=0)
Aug 15 09:55:27 Skynet kernel: Sending NMI from CPU 10 to CPUs 3:
Aug 15 09:55:27 Skynet kernel: NMI backtrace for cpu 3
Aug 15 09:55:27 Skynet kernel: CPU: 3 PID: 0 Comm: swapper/3 Tainted: G           O      4.19.107-Unraid #1

skynet-diagnostics-20200815-0822.zipsyslog

 

I'm going to run memtest on individual stick...

[JorgeB]

Unlikely to be the CPU, very rare to have a CPU go bad, but it's possible.

[cbr600ds2]

memtest completed with no errors on both sticks in different areas.  Now its saying the disk that it was trying to rebuild has two bad sectors.  I'm guessing that's what's causing the errors so I'm going to run w/ that disk emulated and then try to rebuild with a new drive.  

[cbr600ds2]

marked it as solved since it'll take a week or so to get a new drive.