I have two MikroTik x86 routers running as KVM’s on Ubuntu connecting two separate VLAN’s. VLAN10 is 10.0.0.0/24 and VLAN100 is 10.0.55.0/24. The routers are using VRRP on the two VLAN interfaces and a public VLAN. I’m doing NAT on the public VLAN only. These redundant routers are the default gateways for all machines on the VLAN’s.
I am having trouble copying a 114GB file from a Windows Server machine on VLAN10 to a Windows Server machine on VLAN100. It times out after about a period of time (I think it is about 30-45 minutes although I’m not 100% sure since I haven’t monitored it the entire copy time).
At first I had NAT between VLAN10 and VLAN100. I removed that since I read in the forums that could create issues with file copies although I’ve never had that issue with other firewalls. The problem is the same without the NAT.
The Windows Server on VLAN10 (originating server) is reporting an I/O device error on the destination in VLAN100. If I connect the machine on VLAN100, then the file copies without issue.
I’ve tested with making each router active (different attempts obviously). It doesn’t matter which is active, the problem still occurs. I have not tried it by taking a router out of the VRRP mix.
I haven’t run a memtest since it requires a reboot but can probably schedule it. I am not having any other issues with the routers, though, so it would seem like it would manifest itself at other times too, right?
Yes, a memory problem would probably cause other issues, too… but… 30-45 minutes is a really strange timeframe for a file copy to fail. That means it’s been working fine for 30-45 minutes.
I pulled some more statistics. The copy always fails at 58.6% of the copy. The file size is reported at 114.5GB. A 61GB copy of a file in the same directory had no issue. This means the file copy is consistently failing at around 68GB. That number doesn’t make any sense to me. But it definitely appears to be related to the file size since everything else copies fine. Since this is Windows, the file is allocated at the destination prior to the data copy and I’ve verified there is enough space. And the issue only presents itself when copying across the routers. Very odd…any other ideas?
As would be expected, the issue is not with the MikroTik router. Doing some additional testing and verification, I found that this particular file had a corruption issue which was causing the problem. I found this out when attempting to move it to a USB drive.
So, no issues with my MikroTik setup after all. Sorry for the confusion and bad posts.