System specs: 2x 2670 / 64gb ram GTX 970 / GTX 660 / 4x 160gb SSD / 1x NVME 512gb ssd / MOBO with dual NIC

 

My other unraid system I detached from the network to rool out any netbios name conflicts (Both are named Tower..)

 

I followed the instructions and used a proper quality USB drive, altough after booting I have the following problem:

emhttp does noet start. So I have no Webui.

Second problem is: if i choose start unraid with GUI from the bootmenu this does not start, which is strange in itself.

 

Version 6.3.2:

I DOES connect to internet, and I can use NFS/SAMBA on it so it looks like its not a network problem.

Version 6.3.1: 

Doesnt even get a DHCP lease, and wont start GUI either (with 'top' no emhttp entry)

Version 6.2.4 (The version I am running on a seperate system as well without problem)

Same behaviour as 6.3.2

 

I attached the syslog (6.3.2), any help would be appreciated!

 

 

syslog

It looks like you are using a dual CPU MB.  There have been problems with some of these boards not wanting to boot.  So detailed information on the MB would be a good thing to post up.  Some folks have fixed their problem by updating the BIOS.

 

A picture of the console screen can be a help in figuring out where things are hanging.  Just be sure it in focus, not blurry and no big flash reflections in the middle of the text. 

 

If you have a login prompt on your console monitor, you can get the Diagnostics File by logging in and typing   diagnostics   at the command prompt.  The file will be found on your Flash Drive.  Upload the file(s) with your next post with an explanation of what the state of the server was.

 

You do not want two servers with the same name on your network.  That will give you problems as a lot of smb operations involve its use. It can also confuse your router if you attempt to reach the server with the GUI using the server name rather than its IP address.  (BTW, the IP address will always work do you might want to try using that instead of the server name.   i.e.       http://192.168.1.X)

Thanks Frank for helping me!

 

I now you don't want the same names on one network, hence I disconnected my existsing server temporary from the network to make sure no conflicts could arise.

BTW: I tried both direct IP and Name to connect to the webui!

 

 

My motherboard is a asus Z9PA-D8 (VERY nice little atx board!)

I just updated the BIOS to the latest version,  but this made things worse. It wont even get a DHCP lease now?!

 

I'm not sure what you want with a picture of de login screen? Everything it does is in the syslog file is it not?

 

I attached the diagnostics file!

 

Thanks again for the trouble.

tower-diagnostics-20170323-2154.zip

I noticed that the output of the ethtool indicated that a link was not detected which indicates a networking issue.   You (apparently) have three Ethernet ports on this MB.  One is for server management and the other two are used to connect the MB board to the network.  I am not much of expert when using a MB so equipped but but I do recall that earlier versions of unRAID would only use one of them.  So be sure that you try both ports with a reboot in between so the DHCP process will get you an IP address.  (The options for two or more ports has expanded in the latest version but you have to be able to get to the GUI to set everything up.)   

 

Second issue that I saw was this in the syslog near the end:

 

Mar 23 21:53:43 Tower kernel: read_file: error 2 opening /boot/config/super.dat
Mar 23 21:53:43 Tower kernel: md: could not read superblock from /boot/config/super.dat

Is this a virgin install from a from the zipped download of unRAID?  If it a virgin install, I would expect this error to occur but if you copy over the config files from an other server, it indicates a problem reading the Flash Drive. 

Network problem solved. I just connected both eth1 and eth0.

 

So after that I noticed it takes a really long time saying: [sde] Attached scsci disk.

So I unplugged all ssd's and the second GPU aswell.

 

Al worked fine! Booted up nicely and webUI was working albeit EXTREMELY slow. Loading the firstpage took 10 minutes. Installing the trial key another 5.

 

Rebooted again, now with ssd's installed, and it worked! Although still: EXTREMELY slow. 

 

So... troubleshooting time again. What the heck is this now? It is not exactly a strange setup I have is it?! Server mobo with server cpu's and memory should work instantly shouldnt it?

 

Thanks in advance.

 

EDIT:

One cpu is pinned to 100% so found the problem yes? But now I need a fix :-)

Edited March 25, 20179 yr by robertoal

Ok, ran Top again:

 

Udevd is the problem. It is stuck at 100% full time..... so anybody?

Post up your diagnostics file.  You should be able to do it in the GUI.  (  'Tools'  >>>  'Diagnostics'.   )  This time it will download it directly to your PC.

Thanks again for your time Frank!

Diagnostics are attached.

 

I hope you can find what causes the problem.

 

I tried killing udevd, but it simply won't die :-)

tower-diagnostics-20170325-1433.zip

 

EDIT: Just found out: starting the array makes everything faster :-) Strange thing is UdevD still uses 1 core 100%

 

Edited March 25, 20179 yr by robertoal

I only see ssd's in the smart reports for your system.  Are you using them for array drives?  If so, I recall that ssd's are NOT recommended for array drives-- only for cache drives.  I don't know if this is causing your issues or not.  There appears to be a lot of errors in last 10% of your syslog.  (I am not a Linux Guru and these errors are something I am not familiar with.)

 

If someone with more knowledge than I have does not respond in the next 24hours, I would suggest at that point to start another thread, with a new title of something like "Slow server- UdevD uses 1 core 100%".  Be sure to include the diagnostics file from this post and a complete description of all your hardware and any plugins, Docks or VM';s.  And don't be afraid to be very specific in your hardware with manufacturer and model numbers.  Specific exactly which drives are cache, parity and data. 

 

You can't kill udevu because it is a part of the Linux kernel!  As soon as you kill it, it will restart...

 

Edited March 25, 20179 yr by Frank1940

I will follow your advice and start a new thread. Most of the error in the syslog are from the ssd's not being formatted correctly, but after formatting everything was fine.

 

I will put all the ssd's into the cache pool then! Thanks for the tip!

 

Well that makes sense then :-) I will try another clean restart, maybe Udev will then behave normally.

 

Thanks again for your time.