SD-Wan connectivity issues

[from_afar]

We have a single Spoke and Hub SD-Wan set up with only the one spoke and the hub. The hub users connect to Windows machines on the hub LAN to work (open files to work on them in various programs, group policy, printers are all on these servers) and there seems to be connectivity issues sometimes. Some programs crash. Some print jobs crash or take a long time or duplicate. Some of the errors I've seen point to connectivity issues with the server on the hub.

The only events I see in the Spoke's event logs are DHCP events and SNORT update, nothing about connection errors. 

Any ideas where I could look or what might be causing these issues?

Thanks.

[cmr]

Is there any packet loss on the links between the MXs?

[from_afar]

Not that I can see. Would that show in the event logs? I see "sent unexpected packets" every once in a while but nothing about packet loss in the event logs.

[cmr]

It is on the appliance monitoring page, under the uplink tab.  If you scroll down, you should have three graphs like below where minimal packet loss can be seen:

[cmr]

If you go to Monitor/VPN Status and click in the whitespace to the left of the other MX you get to a page that looks like this and it shows end to end stats for the links:

[from_afar]

Thanks very much. I'm limited in what I can see/admin as I'm going through ATT for the service, but I can see the appliance uplink page. The Hub shows no loss, the spoke shows a couple of tiny (<1%) spikes over the last day. The Latency for the hub is steady around 3.5ms and spoke 8ms.

Here is the spoke appliance uplink graph over the last day:

are those tiny loss spikes enough you think to cause these kinds of issues?

[cmr]

@from_afar no that shouldn't cause the kind of problems you are seeing, do you not have the VPN status menu option?

[from_afar]

Thanks for the reply.

I can't seem to get to a view like yours, but I do have a VPN Status menu item under the Organization and Security & SD-WAN menu items. They both seem to take me to the same place:

If I then click to the left of the Description on the networks, it brings me to a page that shows the same info you are seeing. Here is the Hub:

and the spoke:

both from over the last 2 hours. I just don't seem to have a way to get to a page where they all show at once and I can flip between them like you do. I do have the option on the top-right of some pages to "View old version", but I tried that and it didn't show anything like your image--just a different VPN status dashboard.

Again doesn't seem too bad...and the charts seem to show plenty of bandwidth so not sure why the performance for the people at the spoke location seems so bad. 

[cmr]

@from_afar do you only have a single uplink at each site, if so I think you don't get the screen I am showing.  You need to be on an MX with two uplinks for the click to take you to that screen, don't ask me why...

[from_afar]

Ah, right, that makes sense, thanks. At any rate, doesn't seem like there is much interference or anything. Any other ideas of where I could look?

Thanks.

[cmr]

If a user is at the main site, does it all work smoothly?

[from_afar]

Yes, things seem to be working fine at the main site--it's just the SD-Wan users that are experiencing issues. 

[from_afar]

Sorry to jump back in here, but another issue has popped up that is probably the cause. For some reason, the users at the SD-Wan location (and if I VPN into that location I see the same thing) are only getting 3-4Mbps file copy speeds MAX. I've turned off all AMP/IDP/Umbrella settings at both locations (spoke and hub) but not matter what I try, I can't get more than 3-4Mbps (usually hovers around 1.5Mbps after starting up). Sometimes the copy will kick off a bit higher like 8-12Mbps for a couple seconds but without fail, it always almost immediately then slides down to 1-2Mbps and hovers around there:

Any idea how I can find and fix the Bottleneck?

Running MX95 at hub and MX68 at spoke. Both have 100/100 internet. MX are all up to date (MX 18.208). Also running MS-125 and MS-425 switches connected via SPF+ 10G and 40GB cables. I my speeds on the LAN are fine; it's just the SD-Wan spoke that is suffering.

I've also played around with MTU as suggested in some other posts but that has had no effect, either. At a total loss here...

Any ideas would be greatly appreciated. 

[cmr]

You are getting about 50% of performance as 1.8MB/s is about 18Mb/s once you take account of overheads.

I'd try 18.107.10 as I had issues with 18.2xx

[from_afar]

Thanks for the reply.

Yeah not sure why the speed test is so slow too--all of our connections are 1Gig...I do get the same file transfer speeds via Secure Client connecting directly to the hub too, so not sure why it seems to be such a problem with Windows File Sharing. I found another thread that had recomended settings for remote fileshares (https://learn.microsoft.com/en-us/windows-server/administration/performance-tuning/role/file-server/) but that didn't help either. I'm really at a loss...

[cmr]

Are they uncontended connections or shared?

[from_afar]

Thanks for the reply, but I'm not sure what that means. The ISP connection is purely for the SD-Wan/Meraki connection--Wi-Fi is on its own isolated network and ISP (cable) both at the hub and spoke if that's what you mean. I.e. the only thing using the ISP connections (Private fiber at Hub, Verizon Fiber at spoke) are the Meraki MX-68 and MX-95. 

[from_afar]

Any other ideas would be greatly appreciated.