UDP issues after network outage

[Trond.Olsen]

We're experiencing multiple issues with udp-based applications after network outages. A common problem is that udp tracking sessions (I assume from ALG) in PA for DHCP create issues and clients are unable to attain IP-address. This error must be manually solved by clearing sessions. We've also seen this error for other udp applications.

Question is: Are the configuration steps we can take to alleviate our problems with udp tracking sessions getting invalid in PA?

[S.Cantwell]

Hello there

I am not aware of an ALG for DHCP.   UDP has a global time out of 30 secs, by default.

Here is a screen capture of what DHCP looks like on my FW.

Note the start time and receive time (receive time is when the log was received to the traffic log, which logs at session end)

Session ID

38741

Action

allow

Action Source

from-policy

Application

dhcp

Rule

Internal

Rule UUID

6270e459-1170-4495-a10d-2822cd36b2f5

Session End Reason

aged-out

Category

any

Device SN

IP Protocol

udp

Log Action

Generated Time

2019/10/30 15:00:02

Start Time

2019/10/30 14:59:30

Receive Time

2019/10/30 15:00:02

Elapsed Time(sec)

0

Tunnel Type

N/A

So unless someone has manipulated the global timers, or it is anomoly in the software, 30 secs is reasonable max time a UDP is maintained.

That is different for SIP which has an ALG.

[Trond.Olsen]

Hi

In logs and sessions browser the dhcp sessions between ip gateways (remote locations) and dhcp servers are often long running, up to 1-2 weeks. So I'm pretty sure the dhcp application is dependent on more than the udp timeout. It seems unlikely that we should receive continuous dhcp relay traffic from the gateways.

Our config for dhcp application seems standard also:

Also checked ou firewall timeouts which seems to be default:

Session timeout
TCP default timeout: 3600 secs
TCP session timeout before SYN-ACK received: 5 secs
TCP session timeout before 3-way handshaking: 10 secs
TCP half-closed session timeout: 120 secs
TCP session timeout in TIME_WAIT: 15 secs
TCP session delayed ack timeout: 250 millisecs
TCP session timeout for unverified RST: 30 secs
UDP default timeout: 30 secs
ICMP default timeout: 6 secs
SCTP default timeout: 3600 secs
SCTP timeout before INIT-ACK received: 5 secs
SCTP timeout before COOKIE received: 60 secs
SCTP timeout before SHUTDOWN received: 30 secs
other IP default timeout: 30 secs
Captive Portal session timeout: 30 secs
Session timeout in discard state:
  TCP: 90 secs, UDP: 60 secs, SCTP: 60 secs, other IP protocols: 60 secs

---

[S.Cantwell]

Trond,

Can you show a screen capture showing a DHCP session that has been in your Session Browser for more than 2 minutes?

You may need to do a "show session ID xxxxx" from the CLI

I am not able to confirm your issue.  DHCP is just and only 4 UDP packet exchange.

When I go into my Session Browser and look for DHCP, I get none, meaning, I have zero sessions that are in the session table.

But even when I look at a UDP session (DNS in my example), this is what I see

start time : Thu Oct 31 12:10:36 2019
timeout : 30 sec

So we are missing some details... 😕

Please advise. 

[Trond.Olsen]

Here's an example of traffic currently in session today. These are all from gateways doing DHCP relay (IP information has been removed).