Linux8 min readPublished: 14 Feb 2024Updated: 9 Nov 2026

How to Clear the Chrome DNS Cache

Clear Chrome's internal DNS and socket cache, flush your operating system resolver, and check what the public internet actually sees after a DNS change.

G7Cloud Engineering
Platform team
Share:
Abstract illustration representing how to clear the chrome dns cache.
  • Clear Chrome's DNS cache at chrome://net-internals/#dns. Clearing browsing data does not touch it.
  • Flush the socket pool at chrome://net-internals/#sockets too, or an open connection keeps using the old address.
  • Chrome caches separately from your operating system, so you usually need to flush both.
  • None of this changes what anyone else sees. Their caches expire on the record's TTL, whatever you do locally.
  • Lower a record's TTL to 300 seconds at least 24 hours before a planned migration, so the world updates in minutes rather than hours.

How do I clear the DNS cache in Chrome?

Open chrome://net-internals/#dns and click Clear host cache, then open chrome://net-internals/#sockets and click Flush socket pools. Chrome keeps its own DNS cache (its memory of which server address each domain points to), separate from Windows, macOS or Linux. Clearing your browsing history does not touch it.

That is why a site can resolve correctly in a terminal and still show the old server in Chrome. If you want to know whether a server is reachable at all, rather than which address it has, a terminal speed test answers that.

  1. 1Type chrome://net-internals/#dns into the address bar and press Enter. It will not autocomplete or appear in search suggestions.
  2. 2Click Clear host cache. There is no confirmation; the button just works.
  3. 3Go to chrome://net-internals/#sockets and click Flush socket pools. Without this, connections that are already open keep going to the old address.
  4. 4Close and reopen the tab you were testing in, then reload with Ctrl+Shift+R (Cmd+Shift+R on macOS) to skip the page cache as well.

Step three is the one people skip

Chrome keeps connections open and reuses them. Clearing the host cache does not close a connection that is already talking to the old IP address, so the site keeps loading from the old server and the flush looks like it failed. Flush the socket pools as well, every time.

BrowserHow to clear its DNS cache
Chromechrome://net-internals/#dns then #sockets
Edgeedge://net-internals/#dns then #sockets
Bravebrave://net-internals/#dns then #sockets
Firefoxabout:networking#dns then Clear DNS Cache
SafariNo internal cache: flush macOS instead

Every Chromium-based browser has the same page under its own name.

How do I flush the operating system DNS cache?

Your operating system keeps its own cache too. If a record is still wrong after clearing Chrome, this is the next layer down, and often the one holding the old answer.

Windows
:: Command Prompt or PowerShell, run as Administrator
ipconfig /flushdns

:: Confirm the cache is empty
ipconfig /displaydns

:: PowerShell equivalent
Clear-DnsClientCache
Windows also caches failed lookups, so a lookup that failed while DNS was mid-change stays failed until you flush.
macOS
# Ventura, Sonoma and later
sudo dscacheutil -flushcache; sudo killall -HUP mDNSResponder

# Confirm the resolver restarted
sudo log show --predicate 'process == "mDNSResponder"' --last 1m | tail -5
You need both commands: the first clears the cache, the second restarts the service that holds it.
Linux
# systemd-resolved (Ubuntu 18.04+, most modern distributions)
sudo resolvectl flush-caches
resolvectl statistics

# dnsmasq
sudo systemctl restart dnsmasq

# nscd, on older systems
sudo systemctl restart nscd

# Which resolver are you actually running?
systemctl is-active systemd-resolved dnsmasq nscd 2>/dev/null
Many Linux desktops cache nothing at all. If none of those services is active, there is no system cache to flush.

How do I check what the internet actually sees?

This is the check that matters after a DNS change. Your own machine is the worst place to test from, because it sits behind at least three layers of cache. Ask the authoritative nameservers (the servers that hold the master copy of your records) directly, and compare with a couple of public resolvers.

Checking a record properly
# What your machine currently believes
dig example.com +short

# Ask specific public resolvers, bypassing everything local
dig @1.1.1.1 example.com +short      # Cloudflare
dig @8.8.8.8 example.com +short      # Google

# Ask the authoritative nameservers themselves: the source of record
dig example.com NS +short
dig @ns1.yourprovider.com example.com +short

# Full trace from the root servers down
dig example.com +trace

# How much longer will caches hold the old answer?
dig example.com | grep -A1 'ANSWER SECTION'
The TTL (time to live: how many seconds a cache may keep the answer) counts down in real time, so it tells you how long you still have to wait.

Windows has no dig

Use nslookup example.com 1.1.1.1 instead, or install the BIND tools. In PowerShell, Resolve-DnsName example.com -Server 1.1.1.1 gives similar output with TTLs included.

What you get backMeans
Authoritative server has the new IP, resolvers do notThe change is correct: caches are still holding the old answer until its TTL runs out
Authoritative server still has the old IPThe change never saved, or you edited the wrong zone
Different public resolvers disagreePropagation in progress. Normal, and it sorts itself out
Nothing resolves at allNameserver delegation is broken, which is a different problem

Check row two first. It happens more often than people expect.

Why is my site still showing the old server?

There are more caches in the path than most people realise. If the record is correct everywhere and the site is still unreachable, the problem is the server, not DNS, and the system logs are the place to start. Otherwise, work down these caches in order. Each is independent, and any one of them can keep showing you the old answer.

  1. 1Chrome's own cache and socket pool. Both, as above. A private window is a quick check, though it shares the socket pool.
  2. 2Your operating system's resolver. Flush it with the commands in the previous section.
  3. 3Your router. Most home routers cache DNS and have no flush button. Restart it, or test on mobile data to take it out of the path.
  4. 4Your ISP's resolver. You can only wait for the TTL. Query 1.1.1.1 directly to prove the record itself is correct.
  5. 5A hosts file entry. Easy to forget months later. Check /etc/hosts on macOS and Linux, or C:\Windows\System32\drivers\etc\hosts on Windows. It overrides DNS entirely.
  6. 6A CDN or proxy in front of the site. DNS may be right while the proxy still points at the old server. That is a setting to change, not a cache.
Rule out a hosts file entry
# macOS and Linux
grep -v '^#' /etc/hosts | grep -v '^$'

# Windows, in an Administrator PowerShell
Get-Content C:\Windows\System32\drivers\etc\hosts | Select-String -NotMatch '^#'
A hosts entry added during a migration and never removed causes exactly this, months later, for one person on the team.
Pro Tip

The quickest way to take every local cache out of the picture is to test on your phone with wifi off. Different device, different resolver, different network. If the site is correct there, the record is fine and everything you are seeing is local.

How do I make a DNS change propagate quickly?

You cannot flush anyone else's cache. You can shorten how long caches keep your record before you change it, so the world picks up the new value in minutes instead of hours. It has to happen in advance, and it is the most useful thing you can do before a migration.

  1. 1At least 24 hours before, lower the TTL on the records you will change to 300 seconds. Every resolver has to let the old, long TTL run out before it sees the short one, which is why you need the lead time.
  2. 2Confirm the low TTL is live with dig example.com | grep -A1 'ANSWER SECTION'. The TTL is the number before the record type.
  3. 3Make the change. Caches now hold the old answer for five minutes at most.
  4. 4Keep the old server running for at least 24 hours anyway. Some resolvers ignore short TTLs, and visitors who still get the old address reach a working server and notice nothing.
  5. 5Raise the TTL back to 3600 once things have settled, so visitors are not looking the record up on every visit.

Never switch DNS and shut down the old server together

Propagation is not instant, and some resolvers hold records well past their TTL. Leaving the old server answering for a day costs almost nothing. It is the difference between a quiet migration and a day of errors that only some visitors see.

Moving a whole site involves more than one record: the files, the database, the addresses stored inside it and the TLS certificate all need doing in the right order. We move WordPress sites for you: files, database, email and DNS, and you check the result before anything switches.

Frequently asked questions

Does clearing browsing data clear Chrome's DNS cache?

No. The DNS host cache is separate from cookies, history and cached files, and clearing browsing data leaves it alone. Use chrome://net-internals/#dns, then flush the socket pools at chrome://net-internals/#sockets.

Why does the site still load from the old server after I flushed DNS?

Usually an open connection in Chrome's socket pool, which the host cache flush does not close. Flush the socket pools too. If it persists, the cache is further out: your router, your ISP's resolver, or a forgotten hosts file entry.

How long does DNS propagation actually take?

As long as the record's TTL, counted from when each resolver last looked it up. A TTL of 3600 means up to an hour; 86400 means up to a day. Lowering the TTL to 300 at least 24 hours before a change shrinks the wait to about five minutes.

How do I check DNS from a different location?

Query public resolvers directly: dig @1.1.1.1 example.com and dig @8.8.8.8 example.com. That skips every cache on your machine and network. A phone on mobile data gives you a second, independent answer.

Can I force other people's browsers to update?

No. Every resolver and browser between you and them keeps its own cache and expires it on the TTL it was given. All you can do is set a short TTL in advance and keep the old server answering until the change has settled.

Put this into practice on G7Cloud

Every site runs in its own dedicated container behind ScaleShield, with daily backups that are restore-tested every night. Start on the free plan, no card needed.

About G7Cloud Engineering

Platform team

Articles written by the engineers who build and run G7Cloud: UK managed hosting and the AI Website Builder. We write about what we operate every day: containers, backups, databases, and the small-business websites that run on them.

More about G7Cloud →