The value of SOA for BIND/NAMED (CentOS, Red Hat) IN SOA (
1247490039; The sequence number of the
10800; Update
3600; Retry
604800; Obsolescence through
3600); Negative TTL

Sequence number applies to all data within the zone. We started with the unit, which is quite logical. But many people find it helpful to use an ordinal date, for example “1997102301. This is the date in the format YYYYMMDD ^, where yyyy-year, mm-month, DD-day of the month, and NN-count changes the Bo-nal′nyhdannyh on this day. You cannot change the order of the fields, since only this order leads to an increase in the value of the sequence number when changing the date. This is very important: whatever format is not used, the sequence number must be incremented when updating zone data.

When the secondary DNS server establishes a connection with his master in order to obtain information about the zone, first of all, it asks for a serial number data. If the zone data for the secondary DNS server is less than the sequence number of the data, it is considered that the zonal data server is out of date. In this case, the secondary server receives a new copy of the zone. If there is no server backup is unconditional loading zone. As you might guess, when you change the data on the primary zone’s master server, you should increase the serial number.

The next four fields define different time intervals, the default values are specified in seconds:

Update (refresh)
The refresh interval causes the secondary DNS server, how often you should check the relevance of the information for the zone. So that readers have a sense of how to load that generates this value, we inform that the secondary server for every update makes one request the SOA record for the zone. The selected value, three hours, moderately aggressive. Most people would put up with a half-day delay, waiting for them to become part of a network of workstations. When it comes to daily procedures relating to DNS, you increase the value up to eight hours. If the zone data does not change very often, and all secondary DNS servers removed at great distances (like the root DNS servers), it makes sense to think about the larger significance of, say, the interval in 12:0 am.

Retry (retry)
If when the refresh interval expires, the secondary server could not get through to his master (who may very well not working at this point), he tries to run at regular intervals, as defined by the given value. Under normal circumstances, the recurrence interval to shorter than the refresh interval, but this is not required.

Aging (expire)
If the secondary DNS server cannot connect to the principal within a specified number of seconds, the zone data on the secondary server becomes obsolete. Aging zone means that the secondary server stops responding to queries on this area, because the zone data is so irrelevant, that may not be useful. In fact, this field specifies the time when the data becomes so old that it is better not to use them at all. The aging of the week intervals-a common thing, they can be longer (up to a month) in cases where there are problems communicating with the primary source of information. The aging interval must always be much longer than the update intervals and try again; otherwise, zone data will become stale Yeshe before attempting to update them.

Negative TTL
TTL is a time of life (time to live). This value applies to all the negative responses of DNS servers that are authoritative for the zone.