I have been suffering from a slow down which I thought was due to a new crucial cache ssd but had a power cut and parity check running at 

10.5 MB/sec which is about 10th of its normal speed.

 

My system is showing less ram for about a week and I think the issue with 2 dimm slots for CPU 2 but not sure why or even if it has an affect on parity check and general system being really slow.

 

Does anything stand out in the diagnostics?

 

Many thanks

server-diagnostics-20200114-1815.zip

Edited January 15, 20206 yr by bally12345

Don't see nothing special on the syslog, but server load is quite high, does it speed up if you stop all VMs/dockers?

5 minutes ago, johnnie.black said:

Don't see nothing special on the syslog, but server load is quite high, does it speed up if you stop all VMs/dockers?

Dockers were stopped and had just vm running. Disabled VM and DOCKER still the same.

 

Can see CPU running around 15-20%

Try rebooting in safe mode, re-start parity check and post new diags.

is there a command I can run to restart into safe mode? Not got physical access at moment

1 minute ago, bally12345 said:

is there a command I can run to restart into safe mode? Not got physical access at moment

Not directly, but if you click on the flash on the `main tab and go down to the syslinux section and switch to the raw mode you can move the ‘menu default’ entry to be the part of the Safe Boot one you want.

13 minutes ago, itimpi said:

Not directly, but if you click on the flash on the `main tab and go down to the syslinux section and switch to the raw mode you can move the ‘menu default’ entry to be the part of the Safe Boot one you want.

Will Unraid OS Safe Mode (no plugins, no GUI) still give me access to webgui?

Yep

3 minutes ago, bally12345 said:

Will Unraid OS Safe Mode (no plugins, no GUI) still give me access to webgui?

That means no GUI on a locally attached monitor.    GUI access over the network via browser is available in all modes.

Still no access, will have to leave it until tomorrow.

I think it might be time for a update to date system. As iv started to see a few issues pop up.

First X5690 upgrade didn't work, 2 memory dimms stopped working now slow performance which is making me think they might a fault with the board. [emoji848]

Sent from my SM-G973F using Tapatalk

Ok my bad, my VPN had disconnected I have access!

 

Still super slow.

 

Total size:12 TB

Elapsed time:2 minutes

Current position:1.06 GB (0.0 %)

Estimated speed:10.1 MB/sec

Estimated finish:13 days, 16 hours, 41 minutes

Sync errors corrected:0

Edited January 14, 20206 yr by bally12345

Ok my bad, my VPN had disconnected I have access!
 
Still super slow.
 
Total size:12 TB
Elapsed time:2 minutes
Current position:1.06 GB (0.0 %)
Estimated speed:10.1 MB/sec
Estimated finish:13 days, 16 hours, 41 minutes
Sync errors corrected:0

So drove down to server, took out 4 sticks of 4GB memory 2 from each cpu so now down to 32GB

Loaded optimial default settings in bios to make sure nothing was changed by accident...

Looks like everything working again. Did quick download check and was getting ~45MB/s in sabnzbd

Could be all that time the issue was bad memory.

I will do a few more checks and post back

Sent from my SM-G973F using Tapatalk

Nope slowed down again!

 

Might have to do testing locally and try restarting network etc too just to rule out everything.

Edited January 14, 20206 yr by bally12345

Does anyone know how I can manually invoke TRIM or check if TRIM has taken place on the SSD and is XFS fine for a 1TB 

server-smart-20200115-1011.zip
SSD: CT1000MX500SSD1 

 

Crucial Chat is pointless

Quote

Shweta M

I am really sorry but we do not recommend the SSd's to be used in any servers.

Quote

Shweta M

https://www.crucial.com/usa/en/ssd-used-to-be-faster-but-has-slowed-down

10:00:35 AM

You need to follow the steps provided in the above link

 

Edited January 15, 20206 yr by bally12345

fstrim -v /mnt/cache

or

fstrim -v /path/to/mount/point

2 minutes ago, bally12345 said:

Does anyone know how I can manually invoke TRIM or check if TRIM has taken place on the SSD and is XFS fine for a 1TB 

Install the Dynamix SSD Trim plugin to run trim at regular scheduled intervals.

If you only have a single drive in the cache then XFS is fine, although if you ever add another drive to the cache it has to be switched to using BTRFS.

Install the Dynamix SSD Trim plugin to run trim at regular scheduled intervals.
If you only have a single drive in the cache then XFS is fine, although if you ever add another drive to the cache it has to be switched to using BTRFS.

I have TRIM to run at 04:00 daily but I don't know if it has actually run on this drive.



Sent from my SM-G973F using Tapatalk

2 minutes ago, bally12345 said:

I have TRIM to run at 04:00 daily but I don't know if it has actually run on this drive

You can always look in the syslog as there should be entries there when trim runs.

5 hours ago, itimpi said:

You can always look in the syslog as there should be entries there when trim runs.

Jan 15 04:00:02 Server root: /etc/libvirt: 19.9 GiB (21366452224 bytes) trimmed on /dev/loop3
Jan 15 04:00:02 Server root: /var/lib/docker: 16.7 GiB (17922183168 bytes) trimmed on /dev/loop2

Looks like it run this morning but should I be seeing more?

I would have expected to also see an entry for /mnt/cache.

 

You might want to run it manually against that path to see what is reported.

6 hours ago, johnnie.black said:

fstrim -v /mnt/cache

or

fstrim -v /path/to/mount/point

root@Server:~# fstrim -v /mnt/cache
fstrim: /mnt/cache: the discard operation is not supported

the path is correct as I can click the cache and see the index at /Main/Browse?dir=/mnt/cache

Trim doesn't work currently on LSI SAS2 models, but even if it did it wouldn't work on that SSD, LSI requires SSDs with deterministic read zeros after trim, so you should connect it to an onboard SATA port instead.

Trim doesn't work currently on LSI SAS2 models, but even if it did it wouldn't work on that SSD, LSI requires SSDs with deterministic read zeros after trim, so you should connect it to an onboard SATA port instead.

I will try dig out molex to sata power cable

Sent from my SM-G973F using Tapatalk

Trim doesn't work currently on LSI SAS2 models, but even if it did it wouldn't work on that SSD, LSI requires SSDs with deterministic read zeros after trim, so you should connect it to an onboard SATA port instead.

650.01GB Trimmed. I think I might have a bent cpu pin on the board which has been causing all my headaches.

Going to try grab a magnifying glass and see if that is causing 2 dimms not to work on cpu and the 2 X5690s I have.

Sent from my SM-G973F using Tapatalk

15 minutes ago, bally12345 said:

I think I might have a bent cpu pin on the board which has been causing all my headaches.

Going to try grab a magnifying glass

Some good close up low angle photos from multiple directions can help. Cellphones can take remarkable macro shots through a good magnifying glass.