1

I thought I’d start a new thread to see if any other SpiceHeads have had this issue. I do have a Unitrends Case open for this, but we’re not really getting anywhere fast and their response is being pretty slow, in my opinion; although I’m sure it’s great normally.

We had a Unitrends 814S Appliance installed about a month or so ago. To start with all was good, then I started getting the error ‘the host appears to be up, but has a network related error (-253)’ The error isn’t on all servers, but 5 particular servers and all Exchange Databases; which are on one of the servers failing at file level backup. The servers are 2 DC’s, 2 Exchange servers; 1 is just a CAS, and the Spiceworks server. There have been two additional servers getting this error, but they have backed up recently.

The appliance has been rebooted, the servers have been rebooted, all agents and appliance have been updated to the latest versions. I even rebooted the switches that connect everything; we have 2 to load balance dual NIC’s on all servers, which worked, for a couple of days, but the error returned and even after repeating the same steps, the errors are still occurring.

I’m hoping someone else has had this and might come up with a suggestion that could fix this for me.

4 Spice ups
2

Did you recently apply an agent update on the servers? Were they working from int ital setup? I occasionally get the same thing if I try to update the server agent from the Unitrends appliance directly but for whatever reason the agent does not install properly. So you would need to uninstall the agent on the servers, reboot then manually install the agent by going to the appliance share and install the agent from the agent installer folder. That’s what has fixed my issues in the past at least when dealing with that same error.

3

The backups were running fine on the original agent, 9.0.0.4 I think it was, then the error started. Unitrends Support got me to manually re-install the agent 9.0.0.3, which didn’t make a difference.

When the appliance was first set-up I had to manually install the agent on 2008 and 2012 servers, due to UAC being enabled. This has since been disabled to allow the agent update via the Appliance, which it appeared to do successfully.

The servers backed up for a couple of days after I initially rebooted the switches; using the 9.0.0.5 agent, but now they fail again. Today a different server; not one of the particular ones suffering with this issue; threw the same error, I restarted the server agent and it backed up OK, so I tried that on one of the particular servers, but it still failed.

I’ll try a manual re-install of the latest agent and see how that goes.

4

Kevin - I’m sorry to hear that you’re having this issue, and that you feel like you’ve had a longer than expected wait time on a response. Do you mind forwarding over your case number so I can look into what may be holding us up on responding?

5

Just uninstalled the agent, rebooted the server, then manually installed the 9.0.0.5 agent from the Appliances’ SMB share. But I’m still getting the same error.

6

Hi Katie,

I’ve sent you a DM with the case ID :slight_smile:

7

Thanks Kevin! I’m working on this now for you.

8

It could be memory pressure causing Windows to close idle network connections if free memory is less than 15%. At least that is what I was told on a recent case.

9

I’ve just checked the affected Windows servers and the 2 DC’s are currently only using 15-20%, the CAS server is 75% and the Exchange server is 95%-100%, but this is by Microsoft design

10

How long does the backup run until it fails?

Ping the appliance and servers as you are backing them up to see if there is a network issue somewhere. I’ve seen this type of error if a server is rebooted (by wsus or some other update) during the backup but it could just be a flaky network connection. Are these virtual or physical?

1 Spice up
11

I’d have to time how long it takes before it fails. It normally makes 4 or 5 attempts before giving the error.

All affected servers are physical and aren’t rebooting during the backup attempts.

1 Spice up
12

oh… so it doesn’t ever seem to even start the backup?

Is there firewalls turned on? Maybe something is blocking the ports? I cant remember right now but once, on our Exchange server, support had to change the port on the server to a different one.

I found this but maybe my problem wasn’t as similar as I thought.

https://unitrends-support.force.com/UnitrendsBackup/s/article/000001404

13

I would suspect something is using or blocking ports 1743-1749. Use nmap, netstat, or telnet to see what’s going on.

14

Nope, the backup doesn’t actually start.

There are firewalls in place on the servers but all have Unitrends exceptions and Netstat shows 1743 as listening on all of them.

15

But is 1745 open and able to send back? That is what I had to change in the .ini file to get mine to work again. I don’t know, I think you need to keep working with support and I hope they respond quick because I wouldn’t be happy missing out on my exchange backups.

16

Kevin - I believe it’s after hours for you (please correct me if I’m wrong), and that our team has a call scheduled during your business hours of 9:00 - 17:00 GMT tomorrow. If you can connect tonight, please private message me the best contact details for you so that we can set this up.

17

According to netstat 1745-1749 isn’t listening so is available for use. AFAIK 1743 is the only one that listens and then 1745-1749 opens up when the backup runs.

As I mentioned, I have a case open. I just thought I’d reach out to SpiceHeads to see if they’ve experienced the same issue and how they resolved it.

18

How many backup jobs are you trying to run at once? Have you tried to space out the backups to run at different time slots (Say servers 1-3 at 7pm, 4-6 at 9pm, Exchange and 7-8 at 11pm etc) or limit the number of active jobs? I’m just wondering if maybe you are pushing too much through your network at once.

19

I’ve spaced them out pretty well. However, the error occurred even when running 1 manual backup with no other backups running.

Thanks to Katie, I got a call back from Kim at Unitrends Support this morning and despite IP ports 1743-49 being available and not blocked by any firewall, she changed the port to 1744 anyway and set ForceIP to NIC0 of the Appliance. Kim said that from her experience, if it’s a -253 or -255 network error, to change the IP port anyway. I hadn’t done this before as all ports were open and available.

I was told that when set to port 1745, it just uses that port, but when set to 1744 it uses any from 1744-1749. We started off 3 manual backups from affected servers and all backups started straight away, way hay!!!

@katie

5 Spice ups
20

Good deal. Sounds like you are back on your way.