Unreliable and Unresponsive

[russell]

A big thanks to @SSD, @kizer, @Squid, @pwm, @ken-ji, all of your contributions have been useful and encouraging to a novice unraid customer.

 

We ran our server overnight on a new off net network with a plain modem for dhcp, no internet, server as WG Master Browser, and all seems stable so far.

 

Will run it for the next few days in production back on our existing modem router on the main network with internet access to see how it performs.

 

Currently running with no plug ins, no apps, no vm's and no parity for now to see if the most important feature (network shares) works fine on the main network. 

 

Then we will introduce Parity and 3 new drives to replace the drives with smart errors, test again, then we will add in syncthing and a w7pro vm for further testing.

[SSD]

Good luck!

[Zangief]

There are quite a few you tube videos re: tips and tricks on setting up UNRAID, KVM's etc.

 

Look for the SpaceInvaderOne videos - easy and straight forward to follow. My UNRAID setup is significantly tuned now compared to when I started out all thanks to @gridrunner

[russell]

So we had Max (server name) stable on a no internet standalone network, very responsive, no drop outs in browsing or web gui.

Put it back onto the main network, problems with browsing and gui immediately - bear in mind the non network config had all our computers on that network

Put in a router to do NAT between our ADSL Modem Router (Draytek) and Max, and hung the rest of the business of the new NAT router - same problems

 

Conclusion at first thought - get Max on the internet and he crashes

 

Im going to try 2 more tests:

1 - Max on Draytek and only one other computer - no other devices

2 - Max and entire network on a new simple d-link modem

 

Any ideas welcomed, thanks

[Squid]

ip collisions?  Bad cable somewhere? (a bad cable anywhere on the network has the ability to crash other ports on switches)

[kizer]

Do you have any machines on your network that might be sharing the same IP as Max?

At one time I set a reserve IP on my Nest Thermostat and accidentally set my IP on my server to the same IP and for the life of me I couldn't figure out why every time I was streaming a video I could no longer remote control my Thermostat. 

 

 

[russell]

IP address is unique. We have tried with both a fixed IP with a reservation, and also auto ip from dhcp. Both have the same problem.

This problem has been present on:

3 different installs of unraid, 3 different NIC chipsets, 2 different MB/CPU combos, bad ram and good ram, some bad hard disks and all good hard disks. Has to be networking collision of sorts, we have replaced everything else several times
Will try this morning with almost everything unplugged

 

[Frank1940]

A few thoughts.

 

First, config your  unRAID server to force its election as the local master:

 

      https://forums.lime-technology.com/topic/50150-unraid-server-release-62-stable-release-available/?page=17&tab=comments#comment-515607

 

 

Second, here is a little utility that you can install on Windows PC's to determine what computer that PC thinks is the Local Master:

 

       https://scottiestech.info/2009/02/14/how-to-determine-the-master-browser-in-a-windows-workgroup/

 

Note that this program was not developed by unRAID or any member of the unRAID community.  You should use due caution when downloading and using it as you would from a Internet source. However, I have downloaded it in the past and have not experienced any issues with it.

 

Third item, be use to install the Local Master plugin that you can find in the first post of this thread:   

 

      https://forums.lime-technology.com/topic/34889-dynamix-v6-plugins/

 

This plugin is an extension of the part of Dynamix that is the GUI for unRAID.  It will tell which computer your server 'thinks' is the Local Master. 

 

Fourth item, use Capital letters for Computer names and Workgroup names.  Windows for these parameters always assumes that they are Capitals letters while Linux honors capitalization.  I am not completely sure how SMB resolves this situation.  (one example of this problem is that windows will treat Txt.tx and txt.txt as one file while Linux will treat it as two files.  Another case that sometimes crops up is when a share is Media and a Linux Plugin (or  Docker) is assigned to write to a share named media.  Only files from one of these two shares will be visible on the network!)

 

A final item.  If you haven't looked through this thread, you might want to do so.  IF you are finding the some of the links at the beginning of the post don't work (The thread was created while the new forum software was being shaken down), continue on through the thread as many of them were corrected in later posts. 

 

      https://forums.lime-technology.com/topic/53172-windows-issues-with-unraid/

 

 

Edited November 24, 2017 by Frank1940

[russell]

Thanks all for the suggestions (thanks @Frank1940 for the latest contribution) I was just logging in to post a new update
(We did the local master stuff yesterday)

ITS ALIVE!

 

We have an old network attached device that sits all by itself, and never looked at and out of site out of mind. It had its own network address, but something in that device is playing havoc with the unraid server. Its a DIV-R video recorder for our security cameras. During isolation testing, when that network segment was unplugged, up came unraid, when plugged in, it collapsed again.

 

I don't need the DIV-R on the network unless we want to look at footage, so its staying unplugged.

 

I will hold my breath for a few days, however I think its licked - if I plug it back in - MAX disappears, if I unplug the DIV-R - MAX comes back up again.

 

[bonienl]

Check MAC addresses of both devices. They must not be the same.

[Jessie]

What motherboard and processor are you using Russell?

I had a similar problem on a system using a gigabyte x170ws-ecc board and a xeon e3-1245 processor.

The machine is fairly busy. It runs a vm with sbs2008, a windows 10 vm, 2 nextcloud dockers, letsencrypt a couple of separate MYSQL instances and some other minor dockers eg privoxy.

 

It would run for a week then just stop solid.  No response. Chased bad memory and all of the obvious problems, finding nothing.

 

Initially I thought I might have pushed the envelope too far running a domain controller from this system with all of the other systems.

 

Then I discovered by accident that there is a hyperthread bug in the xeon e3 v5 and v6 processors.

 

The fix was a bios upgrade on the motherboard.

 

Did that and it has been running trouble free ever since.

 

I just read your full post.  You are indeed running an X170 mb.

 

Upgrade Bios on both sides to Version F22b.  Problem solved.

 

I upgraded via a usb stick and BIOS.  I found it was better to upgrade to BIOS version F20 first.

 

F20 will prevent you going back to F1/F2.

 

After F20, get it to boot, then upgrade to F22b.

Reset the BIOS after the upgrade then set ups.

 

There will be some side effects.

 

The top PCIe slot wont function anymore for vm passthrough. (talking about the small 2x one not the next one down 16x)

 

The top large 16x and the next big pcie 8x will work ok.

 

I had issues with the bottom pcie 2x/4x slot

You can select 2x or 4x for the bottom slot in bios.  There may be some success in that, but I didn't pursue it.

 

I've built a few of these.  Most of them used a usb controller passed through to the VM so that I could hotswap USB backup drives in the VM.  Could no longer use the top and bottom PCIE for this card. 

 

I built another using a e3-1230 xeon.  It won't start without a graphics card in top slot. This might be a problem if you wish to run 2 separate graphics controllers passed through to VMs.

 

With the e3-1245 you can use the onboard graphics for the Unraid boot up screen and use the PCIe for other things.

 

If not using unraid but using a RAID 10, there is a setting you have to turn off in BIOS otherwise it won't recognise the drives.

 

Edited November 26, 2017 by Jessie

[russell]

Thanks to all for suggestions, been stable since my last post, and since taking the DIV-R off the network. Going to put max all back together again with new 2 x 5bay host swap bays for the hard drives, reformat the drives and reload the data.