966402 - relabs.releng.scl3 TFTP flows

Status: RESOLVED FIXED

(Infrastructure & Operations Graveyard :: NetOps: DC ACL Request, task)

Description

[Dustin J. Mitchell [:dustin] (he/him)]

iPXE> chain tftp://10.26.75.5/pxelinux.0
.. times out

I think that means that we need a flow from relabs.releng.scl3 -> admin1{a,b}.private.releng.scl3.mozilla.com:tftp/udp

Really, though, this should have exactly the "normal" config that other releng.scl3 vlans have as far as DHCP and TFTP.

Comment 1

[Arzhel Younsi [:XioNoX]]

There is an existing rule permitting that flow. Can we have access to the source host to troubleshot it? Can you ping admin1?

Comment 2

[Dustin J. Mitchell [:dustin] (he/him)]

If you can get to the OOB for ix-mn-1.relabs.releng.scl3 and boot it to iPXE you should be able to see this.  It may already be booted to iPXE.

Comment 3

[Arzhel Younsi [:XioNoX]]

[[e-mail] ~]# nc -zvu admin1a.private.releng.scl3.mozilla.com 69

[[e-mail]3 ~]$ sudo tcpdump host 10.26.78.29
listening on eth0, link-type EN10MB (Ethernet), capture size 65535 bytes
05:34:01.997430 IP hp1.relabs.releng.scl3.mozilla.com.52426 > admin1a.private.releng.scl3.mozilla.com.tftp:  1 [|tftp]

I suggest checking if the dhcp configuration received is fine as well as the tftp server config.

Comment 4

[Dustin J. Mitchell [:dustin] (he/him)]

Ah, indeed, the migration script didn't update their DHCP scope, although it updated the address.  So they were getting the correct address, but incorrect netmask and default gateway.  I fixed the scopes.