CIFS/SMB and NFS issues

[tessierp]

Based on experience I had with my QNAP unit which died and now using UNRAID I have been experiencing a lot of CIFS/SMB and NFS issues. Initially I was using CIFS/SMB which was unstable so I tried NFS instead. First started out with the "FSTAB" approach which connects at boot time but seeing that I still had disconnection issues I then tried with "AUTOFS" but it is no use. Eventually my shares will be dropped and the only way to fix it is by stopping the service and starting it again.

 

Not much as change in my setup except that I am now using UNRAID. Has anyone else been experiencing CIFS/SMB and NFS issues like this? I'm starting to wonder if it has to do with BTRFS (most here seem to be using XFS for their ARRAY). Or maybe something with my cache. Not really sure.

 

Anyone has a clue?

[ChatNoir]

You post your diagnostics in your next post.

[tessierp]

Sure, here are my diagnostics. Hopefully this helps.

ptr1-nas-1-diagnostics-20211119-1130.zip

[dlandon]

Your network is dropping offline:

Nov 18 22:09:52 PTR1-NAS-1 dhcpcd[2238]: br0: rebinding lease of 192.168.20.50
Nov 18 22:09:56 PTR1-NAS-1 wsdd[12079]: udp_send: Failed to send udp packet with Network is unreachable
Nov 18 22:09:56 PTR1-NAS-1 upsmon[9622]: UPS [[email protected]]: connect failed: Connection failure: Network is unreachable
Nov 18 22:09:57 PTR1-NAS-1 dhcpcd[2238]: br0: probing address 192.168.20.50/23
Nov 18 22:09:58 PTR1-NAS-1 wsdd[12079]: udp_send: Failed to send udp packet with Network is unreachable
### [PREVIOUS LINE REPEATED 4 TIMES] ###
Nov 18 22:10:01 PTR1-NAS-1 upsmon[9622]: UPS [[email protected]]: connect failed: Connection failure: Network is unreachable
Nov 18 22:10:03 PTR1-NAS-1 wsdd[12079]: udp_send: Failed to send udp packet with Network is unreachable
Nov 18 22:10:03 PTR1-NAS-1 dhcpcd[2238]: br0: leased 192.168.20.50 for 7200 seconds
Nov 18 22:10:03 PTR1-NAS-1 dhcpcd[2238]: br0: adding route to 192.168.20.0/23
Nov 18 22:10:03 PTR1-NAS-1 dhcpcd[2238]: br0: adding default route via 192.168.20.1
Nov 18 22:10:03 PTR1-NAS-1 rpcbind[27514]: connect from 192.168.20.30 to getport/addr(mountd)
Nov 18 22:10:03 PTR1-NAS-1 dnsmasq[13177]: reading /etc/resolv.conf
Nov 18 22:10:03 PTR1-NAS-1 dnsmasq[13177]: using nameserver 192.168.20.1#53
Nov 18 22:10:03 PTR1-NAS-1 dnsmasq[13177]: using nameserver 192.168.20.40#53
Nov 18 22:10:03 PTR1-NAS-1 rpcbind[27553]: connect from 192.168.20.32 to getport/addr(mountd)
Nov 18 22:10:04 PTR1-NAS-1 ntpd[2312]: Listen normally on 3 br0 192.168.20.50:123
Nov 18 22:10:04 PTR1-NAS-1 ntpd[2312]: new interface(s) found: waking up resolver
Nov 18 22:10:05 PTR1-NAS-1 rpcbind[27576]: connect from 192.168.20.30 to getport/addr(mountd)
Nov 18 22:10:06 PTR1-NAS-1 upsmon[9622]: Communications with UPS [email protected] established
Nov 18 22:10:10 PTR1-NAS-1 rpcbind[27670]: connect from 192.168.20.32 to getport/addr(mountd)
Nov 18 22:10:15 PTR1-NAS-1 rpcbind[27701]: connect from 192.168.20.30 to getport/addr(mountd)
Nov 18 22:10:20 PTR1-NAS-1 rpcbind[27732]: connect from 192.168.20.32 to getport/addr(mountd)
Nov 18 22:10:25 PTR1-NAS-1 rpcbind[27763]: connect from 192.168.20.30 to getport/addr(mountd)

Causing remote NFS mounts to fail.

 

Are you using Jumbo packets on your network?

 

I'd also recommend that you use the Unassigned Devices Plugin (called UD).  It makes it a lot easier to mount and manage SMB and NFS remote shares,

[tessierp]

7 minutes ago, dlandon said:

Are you using Jumbo packets on your network?

 

I'd also recommend that you use the Unassigned Devices Plugin (called UD).  It makes it a lot easier to mount and manage SMB and NFS remote shares,

 

I'm not entirely sure what you mean by "JUMP PACKETS" and what do you mean by "Unassigned Devices Plugin"? For the plugin, you mean a UNRAID plugin?

 

From what you said, I suspect maybe my cat6 flat cable could be the issue. Actually suspected that for a while HOWEVER, with my QNAP device it was always able to reconnect so I am not sure why it is not the case. It is true my wiring could be the issue but just curious as to way it is unable to reconnect.

[dlandon]

3 minutes ago, tessierp said:

 

I'm not entirely sure what you mean by "JUMP PACKETS" and what do you mean by "Unassigned Devices Plugin"? For the plugin, you mean a UNRAID plugin?

 

From what you said, I suspect maybe my cat6 flat cable could be the issue. Actually suspected that for a while HOWEVER, with my QNAP device it was always able to reconnect so I am not sure why it is not the case. It is true my wiring could be the issue but just curious as to way it is unable to reconnect.

Jumbo frames are a setting on your network devices.  If you don't know what it is, you're not using them.

 

UD is an Unraid plugin to manage disks and devices outside of the array.  This includes SMB and NFS remote shares.

 

[tessierp]

@dlandon Thanks for sharing this information with me.

 

I looked in my list of plugins and I do not have this "UD" plugin you mentioned and since I do not have JUMBO FRAME then something else is causing the issue.

[ChatNoir]

7 hours ago, tessierp said:

"UD" plugin you mentioned

Just to be precise, the proper name is Unassigned Devices. Most people shorten to UD. It has an addon plugin Unassigned Devices Plus to enhance its functionalities.

[dlandon]

9 hours ago, tessierp said:

@dlandon Thanks for sharing this information with me.

 

I looked in my list of plugins and I do not have this "UD" plugin you mentioned and since I do not have JUMBO FRAME then something else is causing the issue.

UD would not be causing your network issues.  It is just a plugin that adds additional functionality, like mounting remote SMB and NFS remote shares.

 

You need to go to 'Apps' which is the Community Applications (CA for short) and search for Unassigned Devices and install it.  You will then see it at the 'Unassigned Devices' tab of the Main tabs.  You can manage your remote mounts from there.

 

Your network issue can be from cabling, a bad switch, a bad NIC, etc.  If your motherboard has multiple NIC ports, try a different port.  Try different cabling.  Try rebooting your network switch.  Check you network switch configuration if it is managed.

 

Edit: I just noticed in your diagnostics you are using eth3.  Switch to the eth0 port.

Edited November 20, 2021 by dlandon

[tessierp]

48 minutes ago, dlandon said:

You need to go to 'Apps' which is the Community Applications (CA for short) and search for Unassigned Devices and install it.  You will then see it at the 'Unassigned Devices' tab of the Main tabs.  You can manage your remote mounts from there.

 

Your network issue can be from cabling, a bad switch, a bad NIC, etc.  If your motherboard has multi

 

I verified and both Unassigned Devices and Unassigned Devices Plus aren't installed.

 

As for the bad switch that is a possibility I considered but at the moment I do not have another one to test with. Reason that I am also considering it is, at times, the WebUI for the switch becomes unresponsive (could be just the Software, I know).. But my workstation also experiences disconnect issues at times. Bad CAT6 flat cables are another possibility. But I highly doubt the NIC being bad though... But I guess I could also try and see.

 

If the problem is my switch then that doesn't explain why my QNAP server was able to restore connections. I guess it is more resilient to network issues and allows reconnections where UNRAID has an issue with it?

 

Anyway, I have to do more tests I suppose. Trying a different NIC port on my motherboard is the easiest at the moment. But like I said, it is probably the switch that is going bad.