I have the feeling that something is not working properly on my unraid server. It's like my docker applications are running very slow.

 

My server is an HP DL380 Gen9 server with 128GB ram and 2x 2680 v4 xeon cpu's.

 

I totally agree that it's not the latest CPUs, but it's like there's something making all my dockers run incredibly slow. It loads web pages for sonar, radar relatively slowly.

 

When I update a docker application, it is incredibly slow. Sometimes it fails and I have to install again from "add containers" and other times the unraid web gui freezes the page so I can't do anything but wait for it to finish. A few days ago I changed my docker image from an image to a folder, when I read somewhere that it was a good idea if you ran with zfs.

 

Really hope someone can see what's wrong.

dl380-diagnostics-20240425-1317.zip

Following this as I have the same issue and I have seen at least 2 more users with the same issue. Updating containers sometimes takes more than 15 min and while updating unraid gui is unreachable. Restarting a container takes up to 10 min sometimes and usually comes back with an execution error, but it does restart anyway. And several of my containers runs verry slow compared to before updating beyond 6.12.6.

 

I'll create my own thread here when I have more time to pursue this issue. Just following this in case solutions pop up.

 

Edit: I was previous using docker folder but changed back to image just to see if that would solve it, but it didn't.

Edit2: Not to hijack this thread, but adding my diags as well if someone can spot something we have in common.

tower-diagnostics-20240425-2214.zip

Edited April 25, 20242 yr by strike

Have you had any luck figuring this out? If not, does your download client(s) speed fluctuate from a few kb/s to full speed almost every minute?  

Sounds like your Docker folder/image could have been moved to the array.  Can you confirm they are on SSD/NVMe storage?

According to the diagnostics you have the ’system’ share with files on multiple array drives:

system                            shareUseCache="only"    # Share exists on cache, disk1, disk7, disk8, disk10, disk11, disk13, disk14

Since you have the share configured for mover to take no action they will remain there.   Not sure exactly what files they are.   Normally you would want all 0f that share to be on the cache drive for best performance.

 

also you have the docket.img file size configured to be 60GB which is way more than is typically needed suggesting at some point you increased the size, probably to handle a mis-configured container writing files internally to the image file.

I can confirm that my docker image file and folder (when I used that) has always been and still are on the cache drive and no other drives. I don't know why but yes the system folder was on multiple drives, but there was no files there,just the folder. Looks like some point last year my system share was set to use the array or move files there. I have now deleted the systems folder on all drives except cache to not confuse anyone when I post my next diags.

 

I changed back to docker image about two weeks ago and couldn't remember what size I used earlier when I was using image so I just put 60GB. 

 

I had one thought yesterday that my issue might be network related since my speed on my download clients fluctuate from almost 0 to full speed almost every minute, it has never done that before. So I will try to change the ethernet cable after work today if I can find a new one. Maybe it was a coincidence that all the issues started right after updating from 6.12.6. Since I saw at least 2 users with the exact same issues I didn't think the issue could be on my side. Will report back. 

Brand new ethernet cable didn't work either. One more thing I forgot to mention. Most of my containers won't update through the webui anymore. It won't finish the pull for all the layers, see the attached image. I have to update via portainer then press the apply update button in the unraid webui which will state the container is up to date and pull 0 bytes. Only then will it show as up to date. This also started right after I updated to 6.12.8. @itimpi You can move my posts to this thread if you want so OP can have his thread back