What is the output of "show mod" on your switch?
Why no LACP? it is supposed to improve the failover capability.
It is simple to setup: # ifconfig e0a flowcontrol full #not needed # ifconfig e0b flowcontrol full #not needed vif create lacp -b mac vif17 e0a e0b # vif favor e0a not used in multi or lacp ifconfig vif17 `hostname`-vif17 up netmask 255.255.255.0 broadcast xxx.xx.xxx.255 partner vif18 route add default xxx.xx.xxx.x 1 routed on
On the switch: interface GigabitEthernet6/7 description xxx17[e0a] switchport switchport access vlan xxx switchport mode access no ip address flowcontrol receive on flowcontrol send on spanning-tree portfast channel-group 99 mode active
(the active in the last line defines LACP, setting it to on is straight up etherchannel, i.e. pre 7.2 code)
And also on the switch: interface Port-channel 99 description PortChannel for vif17 switchport switchport access vlan xxx switchport mode access no ip address flowcontrol receive on flowcontrol send on spanning-tree portfast
On 6/13/07, Todd C. Merrill tmerrill@mathworks.com wrote:
I'll try to answer most of the questions posed so far...
For the filer involved in the latest incident:
what is the contents of /etc/rc pertaining to the network?
ifconfig e0a flowcontrol full ifconfig e0b flowcontrol full vif create single vif17 e0a e0b vif favor e0a ifconfig vif17 `hostname`-vif17 up netmask 255.255.255.0 broadcast xxx.xx.xxx.255 partner vif18 route add default xxx.xx.xxx.x 1 routed on
What about the stanza(s) from the switch about the connected interface?
interface GigabitEthernet6/7 description xxx17[e0a] switchport switchport access vlan xxx switchport mode access no ip address flowcontrol receive on flowcontrol send on spanning-tree portfast
I use LACP for my vifs (FAS980/R200 and used to on a FAS6070 until it
was changed to GX).
Our network guys told me we do not use LACP.
Are the VIFs in single mode or multi mode? Maybe try going the other way unless you have bandwidth constraints. We run single here because the interfaces are plugging into different switches.
Single mode (see config info above). We do that because we don't need the bandwidth; we often sniff the port; and we only require the redundancy.
Wondering if the vifs have been created out of multiple interfaces on the same or across multiple switches?
Same switch, different blades.
What does the configuration look like? How many interfaces on the switch? How many switches? Vlans? Etc.... Are there any other interfaces configured on different subnets/networks? Do they become unresponsive also?
Switch interface config shown above. Switch has 4x48 interfaces. One switch involved. 40+ VLANs on the switch in total. None of the other subnets/networks experience the unresponsiveness. And, neither do the other five 6070's on this same switch on the same VLAN on the same subnet.
What kind of switches are you using, and what code rev?
Cisco 6506.
#sh version Cisco Internetwork Operating System Software IOS (tm) s3223_rp Software (s3223_rp-IPBASEK9-M), Version 12.2(18)SXF, RELEASE SOFTWARE (fc1) Technical Support: http://www.cisco.com/techsupport Copyright (c) 1986-2005 by cisco Systems, Inc. Compiled Fri 09-Sep-05 21:36 by ccai Image text-base: 0x40101040, data-base: 0x42CC0000
ROM: System Bootstrap, Version 12.2(17r)SX3, RELEASE SOFTWARE (fc1) BOOTLDR: s3223_rp Software (s3223_rp-IPBASEK9-M), Version 12.2(18)SXF, RELEASE SOFTWARE (fc1)
My first focus would be on the LAN switch fabric- making sure that spanning tree is functioning correctly; if you are using VLANs, that your ISL trunks are correct; this smells like some sort of broadcast storm/loop. You're describing a situation where 'all of a sudden' you get a 100% ping loss and then flapping up and down on the network, that really sounds like a switching issue- as if the spanning tree is converging because something ELSE got added to the segment. Do you have Portfast set on the ports the netapp is connected to? How many interfaces do you have?
No other devices on the same switch, on the same VLAN, on the same subnet experience a similar loss of connectivity. :-\
Network guys say the logs show no spanning tree activity during or just prior to the incident.
OR- perhaps someone is introducing a device with the same IP as the netapp?
Also no duplicate IP entries in the logs either.
Thanks for the ideas...keep them comin'...
Until next time...The MathWorks, Inc. 508-647-7000 x7792 3 Apple Hill Drive, Natick, MA 01760-2098 508-647-7001 FAX tmerrill@mathworks.com http://www.mathworks.com