Forum Discussion
nickamon
Nimbostratus
NimbostratusBig IP FQDN Pool Member Resolution from /etc/hosts
Hi,
I've added entries to the Big IP /etc/hosts file to map custom FQDNs to IP addresses (in an attempt to workaround the restriction of having LTM nodes with the same address). I then created an LTM Pool with a member using the custom FQDN hoping it would resolve to the IP address in the /etc/hosts file but unfortunately this is failing.
The pool member is displaying the error "Unavailable (Enabled) - No records returned". Seems like the pool is only able to auto-populate via direct DNS queries. Is there any way to configure the Big IP to consult the /etc/hosts file first?
Thanks
- Paulius
MVP
You shouldn't have to modify the /etc/hosts file because the F5 has a built in feature in the GUI which I believe is outlined in the following article.
https://my.f5.com/manage/s/article/K13206
I do not believe this addition changes what an FQDN resolves to in a pool and looks like it is always resolved by a DNS query which doesn't check the hosts file. We attempted this previously at the company that I work for and it only seemed to effect none tmsh communication.
nickamonThanks Paulius , I am aware of the GUI option but I'm adding this via automation and the entries only need to exist temporarily (I understand that they will be overwritten by updates from the GUI). Disappointedly, I think you are correct, the FQDN pool members are only resolved via a DNS query and don't reference /etc/hosts at all.
Back to the drawing board with this.
Thanks!
boneyardthe article mentioned exactly states that indeed:
Note: Beginning in version 11.6.0, the BIG-IP system uses DNS to resolve host names for nodes and pool members that are identified with fully qualified domain names (FQDNs) and does not consult the /etc/hosts file.
why do you need multiple nodes with the same IP address?
- nickamon
I'm working on migrating our manual configuration of the Big IP virtual servers/pools to an AS3-based automation solution. Our automation takes as input the required information to create a virtual server, pool and all our custom configuration. The problem is that we don't want to first delete the manual configuration, we want to have both configurations running side-by-side so we can test. We have been able to assign a unique VIP for the new configuration but the traffic is destined for the same destination as the manual configuration. When the automation attempts to create a new Pool object with the same members, we run into the error about duplicate nodes.
Using /etc/hosts was an attempt to create a custom FQDN that resolves to the same destination IP address of an existing node to workaround this restriction.
Hope this helps. Thanks!
- boneyard
Ah, yeah that makes sense.
Not sure how side-by-side you want it, else a different BIG-IP might be a way. A lab or trial license might be enough.
Else you could look into route domains perhaps with partitions to make something work, but that will require some designing if you actually want to test with traffic.
