HP5500 HI 48G - Lost connection when a device is rebooted

[it_ejvnior]

Hi folks, good morning.

I'm here again to describe the following situation.

As I said in other topics, I have some HP 1910 48G connected through a 2 Gb uplink in the 5500 HI, (which is the core switch). The core switch behaviour is totally odd when I put a server connected to it to do a reboot or powering off the device.

The reboot process of any machine connected to the core switch makes it inaccessible until the reboot process is finished. If I turn off a machine, the switch stops to respond until the machine is powered on again. I never saw a thing like that.

The STP protocol is disabled in all switches.

Please, anyone could help me to track down this issue? I will really appreciate any help to solve this.

Again, thanks in advance to all HPE community.

[johnk3r]

Hi. You can post the switch logs here?

[it_ejvnior]

More than 10 pages of arp duplicated address when I put the domain controller to reboot.

Mar 28 19:37:58:294 2016 ARP Notification ARP_DUPLICATE_IPADDR_DETECT Detected an IP address conflict. The device with MAC address 74d4-35f0-d5cc connected to Bridge-Aggregation4 in VLAN 1 and the device with MAC address 74d4-35f0-d5cc connected to GigabitEthernet1/0/26 in VLAN 1 are using the same IP address 192.168.11.76.
Mar 28 19:37:18:238 2016 ARP Notification ARP_DUPLICATE_IPADDR_DETECT Detected an IP address conflict. The device with MAC address 0015-1773-4b0a connected to Bridge-Aggregation7 in VLAN 1 and the device with MAC address 0015-1773-4b0a connected to GigabitEthernet1/0/2 in VLAN 1 are using the same IP address 192.168.11.240.
Mar 28 19:37:12:533 2016 ARP Notification ARP_DUPLICATE_IPADDR_DETECT Detected an IP address conflict. The device with MAC address c067-af70-fe3a connected to Bridge-Aggregation5 in VLAN 1 and the device with MAC address c067-af70-fe3a connected to GigabitEthernet1/0/2 in VLAN 1 are using the same IP address 192.168.11.127.

I also captured with wireshark all the reboot process.

Do you want to see the result?

Thanks in advance.

[16again]

Seems like there's some looping going on, wireshark trace might show broadcast traffic going through the roof.
All duplicates are on BAGG interfaces.  Double check if they're set up correctly on both sides.

[it_ejvnior]

Thank you for the reply. In the link aggregation status all Baggs created show running as normal, without any advice. All Baggs was created as static. The Wireshark capture shows a lot of broadcast traffic when the machine is on the restart process. Another way to create a bagg is a dynamic bagg using LACP. Should I try this method for the uplinks or there is another configuration that I've missed during the configuration? All transceivers are HP original acquired accordingly to the data sheet. Again, many thanks for the reply. I will analyze some of the broadcast traffic generated and will do a research about the static Baggs. Thank you 16again.

Regards.

[johnk3r · Accepted solution]

I think the dynamic mode is recommended. You have STP enabled on switches?

[it_ejvnior]

Hi.
No, I do not have the spanning tree protocol enabled.
I read the 5500 HI manual to create a properly dynamic aggregate uplink and I have a doubt about the priority.
Do I need to set up the priority on the core switch with a smaller value than the access switch? As an example, LACP prority on the core switch as 1 and on the access switch as the default value (32768)?
Or simple just the default values on both switches and let the protocol works automatically with the priority?
Thanks in advance.

Edit#1: I changed all bagg types to dynamic. All works flawlessly except one switch. When I put a machine to reboot the lan stops and when I remove the fibre cords from the transceiver the lan goes back to work. I do not know whether is the appliance or some configuration missed. This is the last step to fix everything. Thanks again.

Edit#2: I did the tests switch after switch. One by one.

[johnk3r]

Hi.
You can copy here the log switch?

[it_ejvnior]

Some machines does not receiving LAN or internet. I'm getting general ping failure.
I'm plugged directly in the switch and everything is working. Please, help me.
Edit#1: I activate the dhcp relay on some switches which was not active and powered on the spanning tree protocol with the core switch as root and primary in instance 0. At least the machines are working again. I let you know in the end of the day the results.
Thanks.
Edit#2: In a moment of peace, I turned off all STP from all switches to do a test and now the environment is uprunning without STP.  Everything is fine on all machines but I just realized that the switches does not carrying tagged traffic. An example is the access point traffic for the mobile connections. The untagged connection (vlan 1) works fine using wireless but the tagged (vlan 202) does not work. It connects but not forward the traffic. All ports of the bundled (the bagg interface and the ports) are in trunked mode and the vlans are speciefied.
Edit#3: Is it right?
[Bridge-Aggregation4]display this
#
interface Bridge-Aggregation4
port link-type trunk
port trunk permit vlan 1 100 to 102 200 to 203
link-aggregation mode dynamic
#
return
Edit#4: Tomorrow morning I will create the baggs again accordingly with a HP advice. The first thing, I will remove all the bagg interfaces and check each port status before start to create. The steps are:

First step is
system-view
interface bridge-aggregation 4
With the bagg4 created let's gonna select it and continue with the configurations
link-aggregation mode dynamic
quit
And the next thing to do is select the interface to aggregate in the bridge-aggregation group
interface gigabitethernet 1/0/49
port link-aggregation group 4
quit
And select the other interface
interface gigabitethernet 1/0/50
port link-aggregation group 4
quit
And now return to the bridge-aggregation group interface and configure as a trunk port
interface bridge-aggregation 4
port link-type trunk
port trunk permit vlan 100 101 102 200 201 202 203
And now check the status of the bridge aggregation
display this
End
I pasted here from my notepad. Thanks again and tomorrow I do a new report to the community.
Thanks again.

Edit #5: Problem was solved. All bagg working in dynamic mode. All ports was checked the configuration and all access switches are equal. Now I can reboot the machines plugged in the core and the LAN does not stop to respond. Thanks to all involved.