Bouncing admin ports

[danmunns]

Hi all

My network guy is complaining that the 3PAR admin ports keep resetting themselves on the Juniper switch. Each port is connected as per best practices, and hasn't been set up as a LACP group, or anything that might cause an issue.

Also, nothing shows up in the eventlog, and the port stats seem happy. In fact the 3PAR is blissfully unaware of any issues...

IP Address Netmask/PrefixLen Nodes Active Speed Duplex AutoNeg Status
10.169.112.198 255.255.255.192 0123 0 1000 Full Yes Active

IP Address: 10.169.112.198 Netmask 255.255.255.192
Assigned to nodes: 0123
Connected through node 0
Status: Active

Admin interface on node 0
MAC Address: B4:B5:2F:A7:3A:B2
RX Packets: 21687739 TX Packets: 28804605
RX Bytes: 1577809661 TX Bytes: 23372085791
RX Errors: 0 TX Errors: 0
RX Dropped: 0 TX Dropped: 0
RX FIFO Errors: 0 TX FIFO Errors: 0
RX Frame Errors: 0 TX Collisions: 0
RX Multicast: 2 TX Carrier Errors: 0
RX Compressed: 0 TX Compressed: 0

Could this be something to do with the monitoring of the vSP? 3PAR is running 3.1.2-MU3 (with latest patches), vSP is SP-4.1.0.GA-97.

Has anyone else seen this?

[Cleanur]

It's intentional and caused by the Service Processor periodically collecting status of the array. As part of that process it calls checkhealth which toggles the ports online/offline to check their status before the data is sent home to HP.

[Richard Siemers]

It's intentional and caused by the Service Processor periodically collecting status of the array. As part of that process it calls checkhealth which toggles the ports online/offline to check their status before the data is sent home to HP.

Confirmed. Its part of the near hourly heath check of the standby ethernet port. I can confirm this happens on our 7200 Inserv 3.1.2, but it does not happen with our T-800 running 2.3.1 or 3.1.1

This is probably what he is seeing in his switch logs every hour +/- 5 minutes:


2013 Dec 15 00:50:37.071 %ETHPORT-5-IF_DOWN_LINK_FAILURE: Interface Ethernet113/1/6 is down (Link failure)
2013 Dec 15 00:50:39.985 %ETHPORT-5-SPEED: Interface Ethernet113/1/6, operational speed changed to 1 Gbps
2013 Dec 15 00:50:39.985 %ETHPORT-5-IF_DUPLEX: Interface Ethernet113/1/6, operational duplex mode changed to Full
2013 Dec 15 00:50:39.985 %ETHPORT-5-IF_RX_FLOW_CONTROL: Interface Ethernet113/1/6, operational Receive Flow Control state changed to off
2013 Dec 15 00:50:39.985 %ETHPORT-5-IF_TX_FLOW_CONTROL: Interface Ethernet113/1/6, operational Transmit Flow Control state changed to on
2013 Dec 15 00:50:40.242 %ETHPORT-5-IF_UP: Interface Ethernet113/1/6 is up in mode access
2013 Dec 15 00:50:43.323 %ETHPORT-5-IF_DOWN_LINK_FAILURE: Interface Ethernet113/1/6 is down (Link failure)
2013 Dec 15 00:50:46.231 %ETHPORT-5-SPEED: Interface Ethernet113/1/6, operational speed changed to 1 Gbps
2013 Dec 15 00:50:46.231 %ETHPORT-5-IF_DUPLEX: Interface Ethernet113/1/6, operational duplex mode changed to Full
2013 Dec 15 00:50:46.231 %ETHPORT-5-IF_RX_FLOW_CONTROL: Interface Ethernet113/1/6, operational Receive Flow Control state changed to off
2013 Dec 15 00:50:46.231 %ETHPORT-5-IF_TX_FLOW_CONTROL: Interface Ethernet113/1/6, operational Transmit Flow Control state changed to on
2013 Dec 15 00:50:46.603 %ETHPORT-5-IF_UP: Interface Ethernet113/1/6 is up in mode access

[Andy]

Opened a ticket on this when I first got the 3par. Took a couple of weeks but was finally told by HP that this is by design. I excluded those ports from monitoring and ignore the log entries. I do wish it did not work this way, but some battles are not worth fighting.

-Andy