You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I'm facing a tricky issue with alternc-slavedns when moving a domain from one alternc (A) to another one (B), having A > B when listing config files in /etc/alternc/slavedns.
During a domain migration, if the domain is created on instance B with gesdns=1 (before being removed from instance A), then the next run of alternc-slavedns will read the domain listing for B, but will not create the bind configuration as it finds out that it already exists with master A (which is fine, and we get a notification out of that).
But later, when the domain is removed from instance A, the next run of alternc-slavedns will remove the bind configuration for the domain. As the domain listing of B has not changed, the master is not changed to B, and we won't get notified.
I can fix the situation by running alternc-slavedns -f to force the fetching of all domain listings and add the missing domains in the configuration.
I know this is wrong to have two authoritative name servers for the same domain, but some of our users, which are admin of their alternc, don't or are not aware of that situation.
Our current solution is to periodically monitor the files in CACHEDIR to find if a domain there is not present in bind config. We cal also detect duplicated domains using those list.
The text was updated successfully, but these errors were encountered:
Hi 👋
I'm facing a tricky issue with alternc-slavedns when moving a domain from one alternc (A) to another one (B), having A > B when listing config files in /etc/alternc/slavedns.
During a domain migration, if the domain is created on instance B with gesdns=1 (before being removed from instance A), then the next run of alternc-slavedns will read the domain listing for B, but will not create the bind configuration as it finds out that it already exists with master A (which is fine, and we get a notification out of that).
But later, when the domain is removed from instance A, the next run of alternc-slavedns will remove the bind configuration for the domain. As the domain listing of B has not changed, the master is not changed to B, and we won't get notified.
I can fix the situation by running
alternc-slavedns -f
to force the fetching of all domain listings and add the missing domains in the configuration.I know this is wrong to have two authoritative name servers for the same domain, but some of our users, which are admin of their alternc, don't or are not aware of that situation.
Our current solution is to periodically monitor the files in CACHEDIR to find if a domain there is not present in bind config. We cal also detect duplicated domains using those list.
The text was updated successfully, but these errors were encountered: