| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
This can bring an extra delay, but saves a check in most cases.
|
|
|
|
... and drop the racy code querying dns cache. ๐
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The last one did not make it... Perhaps the cache just needs
a moment to settle?
|
|
This used to crash every now and then with:
> script;error script error: no such item (4)
I guess this is caused by querying the dns cache just exactly when a
record expires. The chance is maximized: The script is started by
scheduler every minute, and the record's ttl is a multiple of a minute.
Let's query records that are not about to expire immediately, and
try again.
|
|
This reverts commit 28da1da49e275fef6089a103edf6c158bbff317f.
Chances are that we have to resolve a CNAME, that does not match when
querying the cache.
How to handle CNAME do multiple A records? ๐คจ
|
|
|
|
|
|
... instead of else-branch.
This makes sure to skip hosts that just became "unknown".
(Possible soon!)
|
|
Just like we do everywhere else.
|
|
|
|
Using `:resolve` we have just one address, but chances are several
records do exist. These end up in cache, so we are happy to find them
there - no need to update then.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
... to prevent confusion after restore.
|
|
... that logs a warning when backup is booted.
|
|
|
|
|
|
|
|
... and pass script name for clean logging.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
... for every script on required RouterOS version.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Christian Hesse