Dynamic DNS page hanging after save
-
Version: 2.6.0-Release (amd64)
After adding a Dynamic DNS Client (in this example, used cloudflare) and pressing save, the page just spins.
If I open a new tab and browse back to the dynamic clients page, it shows what I just added and appears to be functioning normally.
However if I edit the entry and try to save again it just hangs and never recovers. Page loading spins indefinitely.
Anyone else get this?
-
@axsdenied: I'm having similar problems. I've started getting 504 "gateway timeout" errors whenever I try to update my DDNS settings in pfSense.
I'm also having issues just simply getting DDNS to work properly: somehow, pfSense's DDNS service is removing my custom "A" record on my domain, but not replacing/re-adding one (with the correct/new IP address). I've tried posting here, but without any response so far.
-
@superion What DDNS service are you using? Mine did update cloudflare properly and I've adjusted the time table in which it does updates using the Cron app package.
-
@axsdenied: I'm using it with Dreamhost. As far as I can tell from the general logs, it looks like pfSense is issuing a "remove A-record command", which works...but there doesn't seem to be any command issued to add an updated A-record back, which is puzzling. I wish there was some way to look directly at (or better yet, edit) whatever pfSense is doing.
Those 504 gateway-timeouts seem to be another whole issue, however. Recently that has started happening, and I'm not quite sure what is causing that. I peeked around some other forums and someone mentioned it possibly related to the DNS-resolver...but I'm thinking that if this were the case, you would be having general internet trouble. As far as I can see, I don't seem to be having any issues there, so "your mileage may vary" in examining that.
One other place I looked suggested doing a reboot, so I did just that: a "cold" reboot (in the reboot options on pfSense, it's the reroot option, which forces pfSense to cycle-down: stop all packages and services, dismount its drive(s), then remount its drive(s) and restart everything from scratch. That didn't fix things for me, so again, YMMV.
I'll try to follow this forum actively...keep me posted if you find anything out.
-
@axsdenied: any luck? Nothing on my end so far. I'm beginning to think there either something wrong with pfSense OR it may have gotten corrupted somehow (but everything else works...so, not sure about that).
-
@superion Nothing on my end here. But to be clear it still WORKS it just hangs the page.
-
Hello!
I have the same problem on several 2.6 installs. The dyndns Save & Force Update works but will appear to hang. Sometimes it results in an nginx timeout. The rc.dyndns.update script has the same issue.
The problem appears to be in the curl system. The curl_close call at the end of the _update in dyndns.class will hang for 60 seconds before returning.
I am not a curl expert, but this smells like a curl connection cache/pool issue. The dyndns _update creates several nested/overlapping curl sessions. The curl_close in updatedns:_update might be waiting until the shared/pooled/cached connection closes (60sec).
This could also be causing problems with ACB, which also uses curl and can overlap the dyndns update when it is enabled and kicked off by a Save & Force Update.
The easiest workaround might be to tell curl not to share connections, with something like...
if ($this->_dnsService != 'ods') { curl_setopt($ch, CURLOPT_HEADER, 1); curl_setopt($ch, CURLOPT_FORBID_REUSE, 1); //ADD THIS LINE if ($this->_curlProxy == true) {...at the end of the _update function in dyndns.class. This CURLOPT might also help for other curl users (acb, front page widgets, etc...).
Of course, this might break something else and I could be completely off base...:)
John