Parity check crashes system

[Johnyb62]

Found an issue when running parity check on my array after about 10mins caused the system to crash (not just GUI unresponsive but no longer connected to router) .  Booting in "Unraid OS GUI Safe Mode (no plugins)" allowed the parity check to progress so though all was good.  Coming back this morning the server was again unresponsive, physical power down / up required, however after restarting it confirmed the check had completed.  I thought I'd cracked the issue (being one of the plugins was causing it), but seeing as it crashed some point after completing the parity check this cannot be true.  Any ideas?

Thanks

 

[trurl]

Attach diagnostics to your NEXT post in this thread.

[S80_UK]

@Johnyb62 - Replying here since I just saw your post on the Parity Check Tuning thread.

Providing your system's diagnostics could help, as requested above.
But possible causes include...

• not enough RAM
• faulty memory (try several passes of Memtest86)
• possibly overclocked CPU or RAM (easily done if it's a Ryzen CPU since DOCP/XMP defaults are often a CPU memory controller overclock)
• possibly a power supply issue. 

A reliable system should not have problems running parity checks without interruption.

[Johnyb62]

Cannot believe I forgot about this for so long, but system crashed recently so reminded me.

 

Things ive tried:

1. Run Memory test as suggested above, pass - no issues

2. Using different PSU.  As its a Dell with a proprietary PSU connector for the m/board - I used another PSU to power the drives, therefore lightning the load on the main PSU.  But issue still remaind.  Whilst this doesn't completely rule out the PSU, thought maybe the issue was at load.

 

Any suggestions welcomed.  Thanks.

 

ms-diagnostics-20220909-1631.zip

[JorgeB]

Enable the syslog server, start a parity check and post that after a crash.

[Johnyb62]

On 9/14/2022 at 4:48 PM, JorgeB said:

Enable the syslog server, start a parity check and post that after a crash.

Parity check has been running since aprox 8am, server crashed just afer 8pm.

syslog

[trurl]

Connection problems on disk2 but shouldn't cause crash and was hours earlier

[JorgeB]

Nothing logged other than the errors caused by the controller with a SATA port multiplier, those are not recommended but unlikely that is crashing the server, still and since there's nothing logged it points to a hardware issue, PSU or board would be my main suspects, you can also try removing one of DIMMs and test with either one separately.

 

P.S. you should also replace that controller since it will likely causes other issues sooner or later.

 

 

[Johnyb62]

3 hours ago, JorgeB said:

Nothing logged other than the errors caused by the controller with a SATA port multiplier, those are not recommended but unlikely that is crashing the server, still and since there's nothing logged it points to a hardware issue, PSU or board would be my main suspects, you can also try removing one of DIMMs and test with either one separately.

 

P.S. you should also replace that controller since it will likely causes other issues sooner or later.

 

 

Thanks for the feedback, much appreciated.   There we're insufficient SATA ports on the m/b hence why I added the SATA PCI card but its a relatively cheap one as I didn't realise there would be much of a difference in them.  (Desktop PCI-E to SATA 3.0 Controller Expansion Card 4-Port SATA 3.0 6Gbp.)

 

So looks like my options here are:-

1) PSU Issue: - Source another PSU to try replacing current one totally.  (its a Dell m/b so waiting for the adaptor to arrive which allows me to use standard PSU)

2) MB Issue - Think as its a Dell its a proprietary size board not ATX so only option is to seek an identical replacement or bin the PC & build using a suitable MB with a decent amount of SATA connections

3) RAM - I've conducted the RAM test which didn't show any issues, but will try running with a single stick to be sure

4) Replace SATA card - Would welcome recommendations here on what to look for, don't want to purchase another which is not up to the mark

 

 

 

 

 

 

 

[JorgeB]

 

[Johnyb62]

Tried option (1) replacing the PSU, however the issue still remains. 

Whilst the RAM test worked, I may try to swap this out first rather than moving to swapping out the MB / replace the machine.

 

[Johnyb62]

Swapped out RAM, still didn't resolve the issue - so was looking at replacing the MB next.  As it was advised above to replace my controller (SATA port multiplier) I took a punt with this first.  Low & behold this cracked it, was able to run a parity check in one go without a system crash.  Well chuffed.

So it looks like whilst unlikely, it was the dodgy controller card causing the issue.  Thanks to those who posted suggestions.

 

This was the card I went for, based upon the AMS1166 chipset as advised in the recommended controller post.  I dont have a x4 slot on the MB so had to go with the x1.  That said the speed of this thing compared to my older dodgy card is a big jump (parity check went from around 36MB/s to 100MB/s) 

https://www.amazon.co.uk/dp/B097RBLM9G?ref_=pe_27063361_487360311_302_E_DDE_dt_1