Cardlytics
asked on
HP Procurve switch with any other switch plugged into it causes outages.
We have a Procurve 2910 switch.
Any users plugged directly into the switch can operate correctly.
If we plug another switch into any port, any PCs attached to that downstream switch will get an IP and then lose network connectivity. The ARP table doesn't list them at all. Those PCs plugged into the downstream switch can't see any other devices on the Procurve.
I've enabled loop-protect and don't see any loops detected.
I tried disabling spanning tree on the port with the downstream switch plugged in (BPDU filter as well as admin-Edge-port) This made no difference.
I am out of ideas on how to address this.
What can I do to resolve this?
-------------
Note : There are multiple VLANS configured.
---------
vlan 1
name "DEFAULT_VLAN"
untagged 1-46,Trk1
ip address 10.19.0.24 255.255.252.0
exit
vlan 4
name "QA_net"
no ip address
exit
vlan 8
name "Corp-VM-JS"
no ip address
exit
vlan 50
name "QA-VM-JS"
no ip address
exit
vlan 60
name "Bankqa-VM-JS"
no ip address
exit
vlan 14
name "Voice_Net"
qos dscp 011000
tagged 2,4-6,11-14,16,19-20,22,24 ,26-27,30- 31,33-37,4 2,46,Trk1
voice
no ip address
exit
-----
Any users plugged directly into the switch can operate correctly.
If we plug another switch into any port, any PCs attached to that downstream switch will get an IP and then lose network connectivity. The ARP table doesn't list them at all. Those PCs plugged into the downstream switch can't see any other devices on the Procurve.
I've enabled loop-protect and don't see any loops detected.
I tried disabling spanning tree on the port with the downstream switch plugged in (BPDU filter as well as admin-Edge-port) This made no difference.
I am out of ideas on how to address this.
What can I do to resolve this?
-------------
Note : There are multiple VLANS configured.
---------
vlan 1
name "DEFAULT_VLAN"
untagged 1-46,Trk1
ip address 10.19.0.24 255.255.252.0
exit
vlan 4
name "QA_net"
no ip address
exit
vlan 8
name "Corp-VM-JS"
no ip address
exit
vlan 50
name "QA-VM-JS"
no ip address
exit
vlan 60
name "Bankqa-VM-JS"
no ip address
exit
vlan 14
name "Voice_Net"
qos dscp 011000
tagged 2,4-6,11-14,16,19-20,22,24
voice
no ip address
exit
-----
ASKER
I am new to this environment so I am hesitant to remove the VLANs unless they are the clear cause of our issue. I don't believe this is the case because we have 5 other switches in that stack that are similarly set up.
Ports 47/48 are the trunk going to our core switch.
Ports 47/48 are the trunk going to our core switch.
ASKER
The last VLAN 14 is for our telephones. They are set to tag VLAN14. This lets us plug a telephone into any network drop and have it work but still enable QOS and other toys for the VLAN. The "untagged" response should be VLAN1. Then we allow VLAN14 if it is tagged as such.
The VLANs could be the cause of your issue. It would be worth exploring why those VLANs are present in the Procurve switch. Not to mention what routers and routing settings are in use for each VLAN.
Have you tested the other switch while it is connected to one of the untagged ports (1,3,5,7,9-10,15,17-18,21, 23,25,28-2 9,32,38-41 ,43-45) ?
Have you tested the other switch while it is connected to one of the untagged ports (1,3,5,7,9-10,15,17-18,21,
ASKER
@Ultralites
We can plug downstream switches into the rest of our Procurves in the stack. Our problem is unique to this one switch for some reason. Routing is certainly something I've considered. Especially since the ARP table on the Procurve doesn't show the devices in the little switch we plugged in downstream. It may not know where to route stuff because of that. I suspect that if we can find out what is causing that behavior then we may solve the issue. I just don't know where to look. you'd think that if it works for one of our Procurves it should work for all of them unless there is a setting creating this issue. I jst need to find the "setting".
I'll certainly look into the VLANs and remove any not needed for the switches but before I deleted them I'd want to understand how it was causing this issue, if that is indeed the case.
We can plug downstream switches into the rest of our Procurves in the stack. Our problem is unique to this one switch for some reason. Routing is certainly something I've considered. Especially since the ARP table on the Procurve doesn't show the devices in the little switch we plugged in downstream. It may not know where to route stuff because of that. I suspect that if we can find out what is causing that behavior then we may solve the issue. I just don't know where to look. you'd think that if it works for one of our Procurves it should work for all of them unless there is a setting creating this issue. I jst need to find the "setting".
I'll certainly look into the VLANs and remove any not needed for the switches but before I deleted them I'd want to understand how it was causing this issue, if that is indeed the case.
Is the other switch VLAN compatible? If so, VLAN configurations on ports that share the same link must match. So the port you connect to on the other switch will need to have the same VLAN tagging/untagging as the port on the Procurve that you are connecting it to.
The config shown is not the whole story, I guess.
Is gvrp enabled?
Is gvrp enabled?
ASKER
@jburgaard
Thanks for your reply. I'll check gvrp in the morning.
Thanks for your reply. I'll check gvrp in the morning.
ASKER
Sorry for the delay.
@jburgaard : I checked and GVRP is not enabled on our switches.
@ultralites : the switches we are plugging in downstream are just over the counter switches like you'd get at BestBuy. I doubt they support any VLANS. they'll just take the untagged VLAN I'd assume. We can take the same swtich and plug it into the same model Procurve directly above it in the stack and it works fine. The gremlins are unique to the specific switch we are having issues with. There are 5 switches in that stack. Only "switch5" is seeing that issue. Switches1-4 and switch6 are fine.
@jburgaard : I checked and GVRP is not enabled on our switches.
@ultralites : the switches we are plugging in downstream are just over the counter switches like you'd get at BestBuy. I doubt they support any VLANS. they'll just take the untagged VLAN I'd assume. We can take the same swtich and plug it into the same model Procurve directly above it in the stack and it works fine. The gremlins are unique to the specific switch we are having issues with. There are 5 switches in that stack. Only "switch5" is seeing that issue. Switches1-4 and switch6 are fine.
yes, without VLAN support a switch should just see the untagged VLAN.
Do you know which ports on the HP are part of the aggregated link referenced by Trk1?
I have seen many issues with LACP on Procurve switches.
Do you know which ports on the HP are part of the aggregated link referenced by Trk1?
I have seen many issues with LACP on Procurve switches.
ASKER
@ultralites ... ports 47 and 48 are the trunk going to our core switch. Are there specific LACP settings I should take a peek at?
Do you have a default gateway defined on this switch?
ASKER
@ultralites... I will check...
ASKER
@ultralites.... the gateway is set.
I know you have listed the VLANs from your config, to save time on what is enabled, could you post the entire running config?
Feel free to X.x.x.x out any public IPs if you like
sh run
should list your full running configFeel free to X.x.x.x out any public IPs if you like
ASKER
Attached our "show tech" info report from the switch as an attachment, Perhaps it'll spark some brilliance I am overlooking.
switch5.log
switch5.log
As I read the manual the ports in a Trunk should be the same in respect to Fi or Cu
Refer to page 12-7 in the Management and Configuration Guide
http://h20565.www2.hp.com/portal/site/hpsc/template.BINARYPORTLET/public/kb/docDisplay/resource.process/?spf_p.tpst=kbDocDisplay_ws_BI&spf_p.rid_kbDocDisplay=docDisplayResURL&javax.portlet.begCacheTok=com.vignette.cachetoken&spf_p.rst_kbDocDisplay=wsrp-resourceState%3DdocId%253Demr_na-c02564351-3%257CdocLocale%253D&javax.portlet.endCacheTok=com.vignette.cachetoken
Refer to page 12-7 in the Management and Configuration Guide
http://h20565.www2.hp.com/portal/site/hpsc/template.BINARYPORTLET/public/kb/docDisplay/resource.process/?spf_p.tpst=kbDocDisplay_ws_BI&spf_p.rid_kbDocDisplay=docDisplayResURL&javax.portlet.begCacheTok=com.vignette.cachetoken&spf_p.rst_kbDocDisplay=wsrp-resourceState%3DdocId%253Demr_na-c02564351-3%257CdocLocale%253D&javax.portlet.endCacheTok=com.vignette.cachetoken
is there a reason you have IP routing disabled?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Did you try to change the link aggregation, trk1 away from cu+fi?
(port 47: 100/1000T , port 48: 1000sx)
(port 47: 100/1000T , port 48: 1000sx)
ASKER
No solution found.
A large number of your Procurve ports are set to Voice_Net VLAN except 1,3,5,7,9-10,15,17-18,21,2