Hey,
Just noticed that some time ago, my dns cluster setup stopped syncing. I’m using Master → Multi-slave setup as described in your doc, and it worked in an unattended manner for many months (years) and at some point after a hestia update it just stopped doing so. Quick diagnosis brought up the fact that now, if hestia config is set to ‘hestia-zone’, the zones do get pushed to the slave server, but they are completely empty. Changing the config to ‘hestia’ fixes this problem and the records are created properly. Could you please quickly explain what happens here? Thanks!
Install on the primary server where zones are added and managed
DNS_CLUSTER_SYSTEM=‘hestia-zone’
And enable the two slave servers. They should still have empty records if you view them through the control panel. This is normal.
You can use dig @ns1.example.com zone.com to check if it’s responding correctly.
I use a configuration of 1 master server with a panel where sites and DNS records are hosted, and 2 slave servers.
It works flawlessly.
Add to the cron job: */2 * * * * sudo /usr/local/hestia/bin/v-update-sys-queue dns-cluster
I haven’t encountered any major problems personally.
If it worked, I wouldn’t be making this thread. The entires are simply not created with the panel set to hestia-zone.
Are server DNS added to the panel using a password or API keys?
Ah, I see now. It was broken as soon as the panel autoupdated to 1.9.5. I appreciate quality releases :))
Using a cluster aswell and don’t have any issues, you need to provide more informations, so we can try to help you.
Hey, as to my understanding, the ability to diagnose where hestia’s scripts get stalled without deep diving is pretty limited, so I only did a surface level investigation. However, here’s what happened. I’m running a master → slave + slave setup and it’s been working fine for a long time. Then, I notice an avalanche of notifications:
Now, I made some dns changes on the master and went to see what’s happening on the slaves → nothing. I cleared some slave records and ran v-sync-dns-cluster on the master and noticed that the records were created on the slave but with no entries. Then I went and put set -x in the script to see what step it gets stalled and saw something like ‘15 destination server unreachable’ with no further logging. I curled the api endpoint and it was working just fine, I checked the keys and they were all fine as well. So, I deleted and readded the slaves through v-delete-remote-dns-host which didn’t want to work at first because of a missing parent directory (../.. nice touch) which I fixed locally. Still didn’t work. Then I tried changing ‘hestia-zone’ to ‘hestia’ on the slaves and it worked without a hitch again, so I left it at that. Does any of that help in any way? :))
