Skip to content

Est. 2011ยทMicrosoft Partner ID 7033487ยทDelivery under 3 minยทSupport 7 days a week

Your vault is empty.

Free Fix Event ID 7062

Event ID 7062: the DNS server sent a packet to itself in a forwarding loop

12 min read Updated October 5, 2026 Windows Server: AD, DNS & Group Policy

Fix it now

Microsoft publishes 7062 as “DNS Server encountered a packet addresses to itself – IP address w.x.y.z. The DNS server should never be sending a packet to itself. This situation usually indicates a configuration error.” It names four places the loop can live, and a fifth cause: a domain delegated to a server that has no zone file for it.

Read the server’s own addresses first, then set the forwarder list explicitly

ipconfig /all
Set-DnsServerForwarder -IPAddress <resolver1>,<resolver2> -PassThru
  1. Compare the forwarder list against the addresses on this machine. If one of its own addresses is there, remove it: in DNS Manager, select the server, open Properties, go to the Forwarders tab, select the address and remove it. Microsoft’s guidance is to stop and restart the DNS service afterwards.
  2. Check the master list of any secondary zone on this server. If the server’s own address is listed as a master, remove it there too.
  3. Check the notify list of any primary zone on this server for the same thing.
  4. Check delegations of subzones. Microsoft names a delegation of a domain or subdomain to a server that has no zone file for it as a cause of this event.

Set-DnsServerForwarder -IPAddress overwrites the whole list and configures the addresses in the order you supply them, so set it rather than editing it.

If external names resolve promptly against this server and no new 7062 appears, you are done. If not, the next section covers how a query ends up back where it started.

Why it happens

A recursive DNS server resolves a name in a fixed order. It answers from a zone it is authoritative for, then from cache, then through a conditional forwarder if one matches the name, then through the general forwarders, and finally through root hints if it is permitted to use them. Every one of those stages holds an address, and every one of them is somewhere a loop can be introduced.

When the server notices that a query it has received came from one of its own addresses, it stops. Microsoft’s published text is blunt about it: the DNS server should never be sending a packet to itself, and the situation usually indicates a configuration error. The packet is discarded, which protects the server from spending itself into the ground, and the query is now dead. On the client side the same failure appears as a timeout.

Microsoft names four places to check, and the order is worth keeping. The forwarders list, because DNS servers should not forward to themselves. The master lists of secondary zones. The notify lists of primary zones. And delegations of subzones. It also documents a second cause behind the same event: a delegation of a domain or subdomain to a server for which there is no zone file, so the parent keeps referring queries to a server that cannot answer them.

Two variants catch experienced administrators, and neither shows the server’s own address anywhere. The first is a forwarder that is a virtual or load-balanced address which can balance back to this node, so the list looks correct and still loops. The second is a pair of servers with conditional forwarders pointing at each other for a zone neither of them hosts, which loops without either address belonging to the server that logs the event.

The server is listed among its own forwarders

You have this one if An address in the forwarder list also appears in this machine’s own address list. Common after a server is cloned or a build document is reused.

  1. In DNS Manager, select the server, open Properties, go to the Forwarders tab, select the server’s own address and remove it. Microsoft’s documented step after that change is to stop and restart the DNS service.
  2. From PowerShell, set the list explicitly instead of editing it: Set-DnsServerForwarder -IPAddress <resolver1>,<resolver2>, which overwrites the existing list and configures the addresses in the order given.
  3. Forward to your own perimeter resolvers or to a chosen upstream, never to another internal server that forwards back here.
  4. Test recursion against this server directly afterwards.

A secondary zone lists this server as its own master

You have this one if The server holds a secondary zone, and its own address appears in that zone’s master list.

  1. Select the secondary zone, open Properties and go to the General tab.
  2. If the server’s own address is listed among the masters, remove it.
  3. Confirm the remaining masters are servers that actually hold a primary copy of that zone.

This is the second of Microsoft’s four documented places, and it is rarely checked, because nobody expects a zone transfer setting to break recursion.

A primary zone notifies this server

You have this one if The server holds a primary zone whose notify list contains its own address.

  1. Select the primary zone, open Properties and go to the Notify settings.
  2. If the server’s own address is listed, remove it.
  3. Check the other primary zones on the same server while you are there, since notify lists are usually built by copying one another.

A delegation points at a server with no zone for it

You have this one if Names in a delegated subzone fail while everything else works, and the delegation names this server.

  1. Microsoft documents this cause explicitly: a delegation of a domain or subdomain to a server for which there is no zone file.
  2. Remove name server records in child domains that point at the server itself where it holds no zone for the child.
  3. Either create a primary or secondary zone for the delegated domain on that server, or have the delegating server change the delegation to point somewhere that can answer.

Two servers forward to each other

You have this one if No server lists its own address anywhere, queries for names neither of them hosts stall, and both log 7062 within seconds of each other.

  1. Map the chain: for each server, which zones it hosts, which conditional forwarders it has, and where its general forwarders point.
  2. Decide which server actually recurses outward, and make every other server forward to that one and no further.
  3. Delete conditional forwarders for zones a server already hosts, since authoritative data makes them pointless.
  4. Set the surviving lists explicitly with Set-DnsServerForwarder rather than editing them, and retest from a client.

Full reference

Microsoft’s four places, in order

Where What to look for How to fix it
Forwarders list The server’s own address Server Properties, Forwarders tab, select and remove, then stop and restart the DNS service
Master lists of secondary zones The server’s own address among the masters Zone Properties, General tab, select the address in the masters section and remove
Notify lists of primary zones The server’s own address Zone Properties, Notify, select and remove
Delegations of subzones A delegation to a server that holds no zone file for the child Remove the name server records, or create the zone, or have the delegating server change the delegation

Root hints are worth checking too, but they are not on Microsoft’s published list for this event. Treat them as a separate check rather than as part of the documented four.

The events around it

Event Status How to read it
7062 Published The server encountered a packet addressed to itself at the named address, and discarded it
6527 Not published Read the entry for the zone it names. Functionally, a zone held as a secondary or stub that has stopped refreshing
1014 Not published A client-side entry. Read it for the name that failed to resolve and the servers that did not answer
5504 Not published Read the entry for the source address it names. A burst alongside 7062 suggests the same broken path

Setting the forwarder list rather than editing it

Parameter What Microsoft says it does
-IPAddress An array of IP addresses of DNS servers where queries are forwarded. Specify the forwarders in the order you want them configured
-UseRootHint $false Prevents the server performing iterative queries: it forwards unresolved queries only to the servers in the list and does not fall back to root hints
-Timeout How long the server waits for a forwarder before moving on
-PassThru Returns the resulting configuration so you can see what you set

Because -IPAddress overwrites the existing list, it is the safer way to fix a loop: you state what the list should be rather than removing one entry from a list you have not fully read. Decide deliberately about root hints at the same time. Leaving the fallback on means a server whose forwarders all fail will start recursing outward on its own, which may be exactly what you want or may be a route your egress filtering does not permit.

Loops with no self-reference

  • A forwarder that is a virtual address in front of a pool containing this node will loop intermittently, and the list will look correct every time you read it.
  • Two servers conditionally forwarding to each other for a zone neither hosts loops without either address being the server’s own.
  • A delegation to a child that forwards straight back to the parent is the same shape, one layer down.
  • In all three cases, draw the chain out rather than reading one server’s configuration. The loop is a property of the set, not of any one machine.

Diagnosing without leaving debug logging on

Packet-level logging on a DNS server captures everything it handles and writes it to disk continuously, which will find the loop and will also fill a volume if you forget about it. Turn it on long enough to catch the pattern, watch the free space while it runs, and turn it off again. In most cases it is not needed: querying each server in the chain directly, one at a time, and comparing which of them answers and which stalls, identifies the loop without generating a single log file.

Forwarders or root hints

Forwarders in most business networks, because you get faster answers and one place to control external resolution. Root hints are the fallback, and on a network with restrictive egress filtering they may not work at all. Whichever you choose, choose it deliberately: a server configured with forwarders that no longer answer, and root hints it cannot reach, times out twice on every external query before failing, and that is a performance problem masquerading as an outage.

Every code this article covers

Code What it points at Source
Event ID 7062 The DNS server encountered a packet addressed to itself at the named address and discarded it. Microsoft names the forwarders list, secondary zone master lists, primary zone notify lists and subzone delegations as the places to check Microsoft Learn
Event ID 6527 Logged for a zone held as a secondary or stub that has stopped refreshing. Not published by the vendor; read the entry for the zone it names not published by the vendor
Event ID 1014 The client-side view of a resolution that timed out with no configured server answering. Not published by the vendor; read it for the name and the servers involved not published by the vendor
Event ID 5504 Logged when the server receives a packet it cannot parse. Not published by the vendor; read it for the source address it names not published by the vendor

Confirm the fix worked

  1. No address in the forwarder list belongs to this server.
  2. No secondary zone on this server lists its own address as a master, and no primary zone notifies it.
  3. An external name resolves against this server directly and promptly rather than after a timeout.
  4. No new 7062 appears in the DNS Server log over a normal working period.
  5. Clients stop reporting resolution timeouts for the names that were failing.

Questions people ask about this

Is there anything to buy here?

No. This is a configuration loop in the resolution chain, and every command needed ships with the DNS Server role.

Where exactly does Microsoft say to look?

Four places: the forwarders list, because DNS servers should not forward to themselves; the master lists of secondary zones; the notify lists of primary zones; and delegations of subzones. It also names a delegation to a server that has no zone file for the delegated domain as a cause.

Should a domain controller forward to itself for redundancy?

No. A forwarder is where the server sends queries it cannot answer, so pointing it at itself guarantees this event. Redundancy comes from listing two or more genuine upstream resolvers.

Neither server lists its own address and both log 7062. What now?

Draw the chain. Either a forwarder is a virtual or load-balanced address that can come back to this node, or two servers are conditionally forwarding to each other for a zone neither of them hosts. The loop is a property of the set rather than of one machine.

How do I prove which server answered?

Query each one explicitly and compare. Testing through the client’s default resolver hides which link in the chain is at fault.

Related error codes

Was this article helpful?

Your feedback helps us improve our documentation.

Related articles

License Error Event ID 20291: DHCP failover partners rejecting each other’s binding updates License Error Error 0x80092013: revocation checking fails because the CRL is unreachable Free Fix Event ID 1129: Group Policy runs before the network is ready at boot License Error DHCP Event ID 14: the scope address pool is exhausted and leases now fail
โ† Back to Knowledge Base