Bug #3330
closed
Load Balancer showing wrong Status when using aliases for the port
Added by System IT about 11 years ago.
Updated over 5 years ago.
Description
Hello,
I have come across the issue of using Load Balancer with Port Alias's and the status not showing properly in the GUI.
Checking using relayctl show summary shows the pool as being active with all 3 hosts.
I have tested changing the name of the Alias to not include _ underscores, but it does not work either (previous bug was using a (.) period would cause the Unknown - relayd not running? to show in the GUI
The only way the Status in the GUI would show is if i put in a single port number instead of an Alias.
I have restarted the relayd several times as well just to be sure.
Files
I confirm that this is the case in the latest version (2.1.5).
Also there seems to be something wrong with monitors when using aliases for ports.
For example I've set up a alias named "web" with ports 80:443 and set up a HTTPS monitor expecting a 200 status.
The problem is that it's using a https request for port 80 and, of course, it replies with a 400 Bad request so port 80 is always down.
Maybe this should be a different issue ?
Just to bump, this is still the case in 2.2.3. If Daniel Onisoru's ports alias issue above hasn't been made into an issue (will check shortly), only workaround I have found thus far (i.e., have relayd keep both LB routes open/up) is to not use either an HTTP or HTTPS monitor, and instead use the relatively simple ICMP monitor (which is annoying as that's why we're using the LB in the first place. Yes, the machine may well be up, but that tells me nothing about the status of Apache...).
- Subject changed from Load Balancer showing wrong Status when using Alias's for the port to Load Balancer showing wrong Status when using aliases for the port
- Status changed from New to Confirmed
- Affected Version changed from 2.1 to All
- Affected Architecture added
- Affected Architecture deleted (
amd64)
Bug still affecting 2.2.6, Pool Status is marked as down whenever Alias is used, even with ICMP as a monitor.
Confirmed still in 2.3_1. Use of ip alias in address field of virtual server cause " causes status "Unknown - relayd not running?"
Also: Changing alias name does not update the alias name in the load balancer virtual pool list.
Use of _ or just alpha in the alias name makes no difference.
Can confirm the bug still exists in pfsense 2.3.2-RELEASE-p1
Indeed, is still there in 2.3.2-RELEASE-p1, is not assigned to anybody unfortunately and I need to do load balancing on 2 RTP servers that initiate connection on 554 and then use any of the 65535 ports.
Can confirm it is still present in 2.4.2-RELEASE-p1 (amd64). This is really sad, the bug exists for more than 4 years now!
- Status changed from Confirmed to Closed
The relayd Load Balancer has been deprecated and removed from 2.5.0
Also available in: Atom
PDF