+91 88 00 563434 sales@webhostingpeople.net

Tweaking the TTL to reduce downtime after a website migration

Tweaking the TTL to reduce downtime after a website migration

Web site migrations are succeeded by DNS adjustments. A nicely deliberate change minimizes downtime and DNS propagation delays that follows each migration. You don’t need to create a panic amongst your prospects when their emails are misplaced through the change or if their web sites are inaccessible from their location.

It’s all the time a greatest follow to keep up the copy of all migrated accounts on the previous host for a couple of extra days in order that the web sites stay on-line for purchasers whose net/mail requests are nonetheless served by the previous server. Their ISP’s caching DNS servers/resolvers nonetheless serve cached DNS data to purchasers even after you’ve gotten pointed the area to a location.

The issue with DNS caching is that it’s the main trigger for delay in propagation of the DNS adjustments (Tackle data, mail change data and so forth.) you made after the migration. This is because of the truth that many of the DNS visitors is eased with DNS caching the place your resolver/caching DNS caches DNS data for a predefined time. The entity that determines this time is TTL( Time to stay). By default the TTL values for many useful resource data are 86400 seconds (24 hours). Any resolver that queries for the A document of your area will cache that data for 24 hours.

To unravel this downside to an extent, you’ll be able to decrease the TTL values of the useful resource data (MX, A) earlier than you alter the IP deal with. We assume that you’ve got root/sudo entry to the authoritative nameserver.

Login to the DNS server and all you need to do is edit the zone file within the authoritative nameserver of the area (You might use DNS zone editor in cPanel if the service is operating on a cPanel server)

Lets open xyz.data’s zone file utilizing vi editor

vi /var/named/xyz.data

These are the primary few traces of the zone file, you’ll be able to see the primary line $TTL that defines the default TTL for all present data to 86400 seconds.


$TTL 86400

xyz.data.       IN    SOA  ns1.instance.data. admin.xyz.com. (

2012080907      ; Serial

10800           ; Refresh

3600            ; Retry

604800          ; Expire

300             ; Destructive Response TTL



We’ll change this to 300 seconds (5 minutes)

$TTL 300

We’ll increment the serial quantity from 2012080907 to 2012080908 in order that the DNS server can confirm the contents of the zone file.

Now reload the DNS service
/and so forth/init.d/named reload

Reloading named:              [  OK  ]

After this we’ll look ahead to the final TTL period of time (24 hours) to guarantee that the caching nameservers all all over the world that has beforehand queried and cached this area’s DNS data has purged the data. As soon as that point has handed we’ll change the A document to the brand new server’s IP. I wouldn’t say that is 100% foolproof as there are caching nameservers which enforces their very own TTL values and net browsers too cache DNS data for a default period of time.

Be aware:- Be sure you revert the TTL worth change later , a low TTL worth will result in a rise in DNS visitors.