So I'll begin with the equipment I have:
Cyberoam CR100ing Netgear XS728T Cisco SG-300 52
The Cyberoam is performing L3 functions between 3 VLANS, 1, 20 and 30.. and up until tonight has been working just fine.
Tonight I had to reboot a couple Windows servers, two on VLAN1, 1 on VLAN20. I also have a Qnap TV-1635 on VLAN30.
After rebooting the servers, and power cycling the SG300 (to enable Jumbo Frames) I lost the ability to ping devices on VLANs 20 and 30 that had static addresses. Those who had pulled DHCP however, were fine. Everything on a specific VLAN can ping it's own GW
Take the Qnap for example.. has two hot interfaces, I could reach the web GUI via the 1GB link that was DHCP. I lost communications with the static'd 10G interface. Nothing I did could resolve it. During troubleshooting I added a 10g nic to a Backup server and static'd it on VLAN30.. couldn't ping anything on VLAN30 but the GW. I could reach devices with static'd addresses from earlier this week on VLANs 1 and 20... but not 30.
I then set the 10g interface on the Qnap to DHCP... would not pull an address. Made sure it's port was untagged for VLAN1, tagged for all others and it's PVID was changed to 1. And every single port on the Sg-300 is untagged VLAN1, tagged all others.
No dice.
I've had so many weird issues with using the Cyberoam for routing that I'm beginning to wonder if I shouldn't enable L3 on the 10G Netgear switch. (which terminates the Qnap, backup server and Vsphere hosts). Or, since this is a small network....say less than 100 devices. Should I cut my losses and just revert it all back to VLAN for simplicities sake? I'm used to have servers, storage, networking, etc all on their own VLANs but those were larger environments. Should I keep it simple stupid?
No comments:
Post a Comment