Upgraded ASAs, now internal communication is unstable/sporadic

We have two ASA 5520s in an active/standby pair. Hanging off of those are two core switches (WS-C3750X-48T-S), each serving two different environments, and each routing multiple vlans. Recently it came to light that we had not updated the ASAs since 8.3(2)44 so there was a huge push to get that done over the weekend. They are now on 8.4(7)31. We did an incremental upgrade:

8.3(2)44 > 8.4(1) > 8.4(5) > 8.4(6) > 8.4(7) > 8.4(7)31

using the zero downtime method.

Now, we are having problems with servers contacting each other within the same core switch, even within the same vlan. If we clear arp, that might fix it for a short time or it might not. In one environment, we had to add an ACL that said:

access-list network2_in extended permit ip object network2 object network2

that allowed servers to talk to each other within the same network. It made it better but did not fix the problem entirely. There is already an any-any on the other network, but it is having the same issues. Email is not going out, because the Exchange server cannot send mail to the smart host appliance in the same network/vlan. If we ping the appliance from the Exchange server, connection is restored temporarily.

Questions:

Why does the ASA have anything at all to do with communications within a core switch, including within a single vlan?

What the crap is going on?
EE.png
LVL 13
Yancey LandrumTechnical Team LeadAsked:
Who is Participating?

Improve company productivity with a Business Account.Sign Up

x
 
Yancey LandrumConnect With a Mentor Technical Team LeadAuthor Commented:
Sorry for the delay; we got it sorted out. It turned out to be proxy arp.

We had some NAT overlap in our config that was causing the ASA to answer ARP requests for addresses on the subnets with its own MAC address. We could have just added no-proxy-arp to the end of some representative NAT statements for each subnet, but ultimate decided to just disable it at the two interfaces since we have no one-to-one address NATs:

sysopt noproxyarp network1
sysopt noproxyarp network2

Here are some links that led us to the answer:

http://www.cisco.com/c/en/us/support/docs/security/adaptive-security-appliance-asa-software/116154-qanda-ASA-00.html

https://supportforums.cisco.com/discussion/12938596/asa-interface-hang-fixed-clearing-arp

https://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=1028373

http://www.cisco.com/c/en/us/td/docs/security/asa/asa93/configuration/general/asa-general-cli/route-overview.html#66617

Thanks for your willingness to chime in.
0
 
skullnobrainsCommented:
If we ping the appliance from the Exchange server, connection is restored temporarily.

that is no actual solution but this suggests you are facing a problem related with a security feature in ASA. most likely the ASA sees the same IP on multiple ports and assumes some spoofing is going on and kills the mac... or some weird interaction with the core routers. i guess that same feature was not enabled by default but now is.

can you post information regarding your setup including the core routers and ASAs ?
0
 
skullnobrainsCommented:
interesting. thanks for sharing.
0
 
Yancey LandrumTechnical Team LeadAuthor Commented:
We solved it ourselves.
0
Question has a verified solution.

Are you are experiencing a similar issue? Get a personalized answer when you ask a related question.

Have a better answer? Share it in a comment.

All Courses

From novice to tech pro — start learning today.