High CPU after upgrading from 6.5.3 to 6.11.1

[fishface]

I have just upgraded from 6.5.3 to 6.11.1 and immediately noticed that my lowly old rig is running at around 50% total load, where as previously it would run at 10-15%, rarely saw it going over 50%.

 

I have no active shares, and only one docker running, miniDLNA.

 

I admit it's an old lowly rig, but has been working well.

AMD Athlon™ II X3 405e @ 2300 MHz, 8GB RAM,  6 drives in total.

Is high CPU use expected after an upgrade and eventually settles down?

 

Or is that my old rig is, well, too old now?

If it's considered old, is it straightforward to rollback to 6.5.3?

I did backup the flash drive before the upgrade.

 

Thanks

Edited October 31, 2022 by fishface

[fishface]

I'm replying to my own topic - after 20mins or so it did indeed settle down to its typical 5-11%!

 

Still checking, but circumstantial evidence seems to indicate I may have taken a performance hit, copying some data around and the transfer speed seems to have taken a nose dive, and CPU is at 28-50% while copy this 60GB of data.
 

I recently moved the same data around, yesterday, and it took several hours, but now it says it will take 2.5 days and is going dog slow.

 

It might be that miniDLNA is re-indexing....maybe.

I'll see how it goes, and report back.

Edited October 31, 2022 by fishface

[fishface]

So, got another issue.

 

I think 6.11.1 might be a bit too much for my lower power system, for the first time ever uses unRAID it failed to stop the array, I hit stop the array, as I had changed some disk settings in UI, and it advises to reboot, so I went to stop the array, and it got stuck at "stopping array", left it for 15mins and still nothing.

 

WebUI was still responsive, and it showed all 3 cores pegged at 100% (another first).

 

shfs is stuck 68% and smbd at 28%

 

So what to do? It's now stuck at "stopping array"

 

Also, under the Dashboard, far right under the Parity information, you have Array information, and it states "Array Stopped" but as I have said, in the lower left corner, in the lower bar, it says "Array Stopping", in orange, and next to that "Stopping services..." - it seems confused.

 

I have hit the "Reboot" button, just before I did that I sent continuous  pings and it's responding, as soon as I hit reboot the pings stopped, as one would expect, but shows that something is happening, however, it did not reboot has uptime is still at 23hrs.

 

So, what next, hard power off using the physical power button?

wait time is very high in top, but the other odd thing is the webUI shows CPUs pegged at 100%, but when usign ssh and htop, there isn't much going on, 2% for shfs, emptd 1% etc.

 

I gave and hard powered off, "Power Off" in UI did nothing, also powerdown -r in command line did nothing.

It all came backup, all the drives are online and report "green", it did detect an unlcean shutdown., but is stuck at "Starting array", should I just wait now?

rutland-syslog-20221101-1629.zip rutland-diagnostics-20221101-1627.zip

rutland-syslog-20221101-2305.zip rutland-diagnostics-20221101-1805.zip

Edited November 1, 2022 by fishface

[JorgeB]

Load average was very high in the first diags, also a lot of SMB related noise, if you don't have Macs disable "Enhanced macOS interoperability" on Settings -> SMB

[fishface]

I tried to change the mac setting, it was on "Yes", when I try to change it after hitting Apply the waving lines come up but never seems to stop, left it for over a minute. So, I click on Dashboard and then go back to settings and it's still set at Yes, it seems like I'm having UI issues, I tried another browser, same result.

 

Can I manually edit the smb.conf?

Also the drive all still have green balls next to them, and I have zero read writes from all the disks, also reporting temps of 43C which is the sort of temps I get when the drives are spinning, so it's like the UI is not refreshing, and the array is still stuck at Starting Array and all the disks are spinning.

 

I left it overnight, wondering if it indeed was misleading, was started doing a parity check and the UI is just misleading me, it was still the same as I left but with one minor difference, in the top right corner the "Parity has completed" pop-up dialog box was shown, it wasn't before, so I do indeed now wonder if it completed the parity check and it's all UI issuses.

Edited November 2, 2022 by fishface

[JorgeB]

Try booting in safe mode to rule out any plugin related issues.

[fishface]

Is there a command line option for that?

 

The UI is acting very strange, as I've pointed out above, every time I've tried "Reboot" or "Shutdown" it does nothing.

 

And...I just clicked on Reboot and checked "Safe mode" and it does nothing, so the UI is messed up.

 

As a side note, had zero issues like this with 6.5.3, in fact no issues at all for years.

 

I see this in the logs

 

Nov  2 16:39:55 rutland nginx: 2022/11/02 16:39:55 [error] 3013#3013: *723526 upstream timed out (110: Connection timed out) while reading response header from upstream, client: 192.168.1.8, server: , request: "POST /webGui/include/Boot.php HTTP/1.1", upstream: "fastcgi://unix:/var/run/php5-fpm.sock", host: "192.168.1.9", referrer: "http://192.168.1.9/Main"

 

This taken 5 mins after I have hit the Reboot, nothing happening.
 

16:32:03 up 22:53,  1 user,  load average: 1.96, 1.56, 1.28

 

At this point I'm wondering if I should rollback to 6.5.3 on the USB stick backup I did.

 

Screenshot below, shows the live output of the console, while pinging as I hit the Reboot button, and it does not reboot.

 

Edited November 3, 2022 by fishface

[JorgeB]

You'll have to force a reboot if needed, then choose safe mode in the boot menu options.

[fishface]

Did a hard power off, then activated Safe Mode, rebooted.The drives showed as off-line in the window under Array in the Dashboard, which is as one would expect.

 

I disable the SMB Mac option, which I was now able to do, previously it wouldn't stick.

 

I then started the array, and it is still the same, stuck at Starting Array, but the drives in the Dashboard view now show online,but no shares are available in the UI as it states "There are no exportable user shares".

 

Also, if the UI is to be believed, and I'm increasingly not trusting it, the disk are never spinning down, they are all constantly in  a green ball state, with zero bytes being moved and nothing happening.
 

So what next?

Edited November 3, 2022 by fishface

[JorgeB]

Are you using Firefox?

[fishface]

I used Firefox, Google Chrome, and Konqueror, all the do the same, also tried another PC, same result.

 

 

[fishface]

It's like the UI grabs a first load of info just after boot, and then it is locked/crashed or something, and parts of it don't update.

[fishface]

Shall I upload a fresh diagnostics?
 

I think the disk are constantly spinning, as the drives temps are on the high side.

Edited November 3, 2022 by fishface

[JorgeB]

Yes, upload new diags.

[fishface]

New diags

rutland-diagnostics-20221103-1101.zip

[JorgeB]

Post a screenshot of the main GUI page, according to the syslog array didn't start.

[fishface]

Here we go

Edited November 3, 2022 by fishface

[JorgeB]

Something weird is going on there, I would suggest backing up the flash drive, then redoing it and just restore super.dat and the key to rule out any config issues.

[fishface]

Just so I have this correct.

 

1. Backup flash drive
2. Use the unRAID USB creator tool on the original USB flash drive, using the backup I just made, as I can't use the "Update" plugins as the plugins are disable
3. Reboot etc

 

Or

1. Backup flash drive
2. Use the unRAID USB creator tool on the original USB flash drive, but manually download a new copy of unRAID6.11.1
3. Reboot etc

 

 

 

Edited November 3, 2022 by fishface

[JorgeB]

Before or after rebooting but before first boot restore the key and super.dat from the backup.

[fishface]

Just checking something, the first screenshot shows the contents of my 6.11.1 broken config directory, the second screenshot is the backup now copied to the original USB thumb drive, with the super.dat and the key.PLUS file copied over from the 6.11.1 broken config directory.

 

My query is this, new USB thumb drive config directory has a lot less content/files - is this expected?

[JorgeB]

Yep.

[fishface]

Trying again, it says the drive unbootable 

 

Forgot the make_bootable

Edited November 3, 2022 by fishface

[fishface]

Ok, got it rebooted, it changed the IP (I had a static IP) so gave me a bit of a heart attack for a few seconds.

 

Could be issues with disk2 and 3, even though they recently passed a SMART last week.

 

And of course, it has lost all my shares configs, but I guess I can copy that over from the backup once the other issues have been fixed.

 

Edited November 3, 2022 by fishface

[JorgeB]

SMART warning are likely about the existing UDMA CRC errors, you can acknowledge them, and start the array to see if it works better.