Skip to content

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

Your vault is empty.

Free Fix 0x80070079

The semaphore timeout period has expired: 0x80070079 copying to a share

11 min read Updated October 5, 2026 Networking, Sharing & Printing

Fix it now

0x80070079 is Win32 121, “The semaphore timeout period has expired”: a wait for a response ran out. During a file copy that means the transfer stalled long enough for the client to give up, and the usual reason is the link underneath rather than the share, the permissions or the files.

Run these on the client before you change anything, and keep the output

Get-NetAdapterAdvancedProperty | Format-Table Name, DisplayName, DisplayValue
Get-NetAdapterStatistics
  1. Copy a small file to the same destination. If small copies work and large ones fail part way, this is link stability and not permissions.
  2. In Device Manager, open the network adapter, go to Power Management and clear the option that lets Windows turn the device off to save power.
  3. On the Advanced tab of the same adapter, turn off the large send offload entries and any receive coalescing setting, one at a time, retesting after each so you know which one mattered.
  4. Update the network driver from the machine or adapter manufacturer rather than a generic package.
  5. Change the cable and the switch port. On wireless, retest wired to take roaming out of the picture entirely.
  6. If the copy fails at the same point in the same file every time, test the disks at both ends. A read error presents exactly like this.

Keep the Get-NetAdapterAdvancedProperty output before you start. Advanced settings differ by adapter and by driver version, and you will want to put them back if none of them was the cause.

If a large copy now completes in both directions, you are done. If not, the next section explains why a small file succeeds and misleads you.

Why it happens

A file copy over the network is one long conversation, not a series of independent requests. The client keeps a session open, sends data continuously and expects the far end to keep acknowledging it. A stall of a few seconds anywhere in that path exhausts the wait, and the copy fails with whatever it had transferred discarded. A small file finishes inside a window where the link happens to be behaving, which is why it succeeds and convinces people the share is fine.

The commonest source of those stalls is the network adapter itself. Offload features hand parts of the work to the adapter’s hardware, and a driver bug in that path produces exactly this pattern: fine under light load, failing under sustained transfer. Power management is second: an adapter allowed to power down mid-transfer, or an energy-efficient link negotiating down while data is in flight, both stall the conversation. Physical faults are third, and a marginal cable or a duplex mismatch will do the same while passing every casual test.

Two of the companion codes are documented and describe the network refusing rather than stalling. 0x80070036 is Win32 54, “The network is busy.” 0x80070037 is Win32 55, “The specified network resource or device is no longer available.” The other two, 0xC00000C9 and 0xC000009A, are status values Microsoft does not publish a meaning for; what they are useful for is telling you the session ended rather than timed out, and that is a different investigation from this one.

Adapter offload features or a driver fault

You have this one if Sustained transfers fail while interactive use is fine, and the machine is otherwise healthy.

  1. Record the current values with Get-NetAdapterAdvancedProperty so you can put them back.
  2. Turn off the large send offload settings and any receive segment coalescing option, then retest with a large copy.
  3. Update the driver from the hardware manufacturer. Generic drivers are frequently the ones with the offload bugs.
  4. Where jumbo frames are enabled, confirm every device in the path agrees on the same size. A mismatch produces this exact symptom.

Change one setting at a time. Turning six things off at once tells you nothing except that something among them mattered.

Power management on the adapter

You have this one if Failures cluster after periods of light use, or on laptops running on battery.

  1. Clear the option allowing the computer to turn the adapter off to save power, in Device Manager under the adapter’s Power Management tab.
  2. Turn off any energy-efficient Ethernet or green Ethernet setting on the Advanced tab.
  3. Check the active Windows power plan is not throttling the adapter, especially on battery.

A physical layer problem

You have this one if The link reports errors, or the fault follows the cable or the port rather than the machine.

  1. Replace the patch cable. It is the cheapest test available and it resolves a surprising share of these.
  2. Move to a different switch port, and check the negotiated speed and duplex match at both ends.
  3. Read the adapter’s error and discard counters with Get-NetAdapterStatistics before and after a failed copy.
  4. On wireless, test the same copy wired before investigating anything else.

The session was torn down at the other end

You have this one if The failure is abrupt rather than a stall, and carries one of the status values rather than the timeout.

  1. Check the server for a restart, a cluster failover or a service restart at the time of the failure.
  2. Across a site link or a VPN, check whether an intermediate firewall drops connections it considers idle. A long single-file transfer can look idle to a device counting packets per flow.
  3. Reconnect the share and retry. A torn-down session does not recover on its own.

The far end has run out of resources

You have this one if A small appliance or an older server fails under several simultaneous transfers and is fine with one.

  1. Reduce concurrency: run one large copy at a time rather than several.
  2. Check what else the device is doing during the copy, including its own scheduled scanning or indexing.
  3. Where a security product scans every file as it is written, test with that scanning temporarily narrowed, with approval, to establish whether it is the limit.
  4. If the device is simply undersized for the work, no client setting will fix it.

Full reference

What the five codes are, and are not

Code Published meaning Source
0x80070079 Win32 121: the semaphore timeout period has expired Microsoft Learn
0x80070036 Win32 54: the network is busy Microsoft Learn
0x80070037 Win32 55: the specified network resource or device is no longer available Microsoft Learn
0xC00000C9 No meaning published Not published by the vendor
0xC000009A No meaning published Not published by the vendor

The two undocumented status values are still useful, but only as a shape. If the copy stopped abruptly rather than hanging, and one of them is what you have, look for something that ended the session – a restart, a failover, a device dropping the connection – rather than for something that slowed it down. Do not build a fix on a meaning nobody publishes.

Reading the failure pattern

Pattern Where to look
Small files fine, large files fail part way Adapter offload, power management or cabling
Fails at the same byte in the same file The source or destination disk, not the network
Fails on wireless, works wired Roaming or signal, not the share
Whole session drops, not a timeout The server restarted, failed over, or a device dropped the connection
Only under load, and the far end is a small appliance Resources on the appliance, including its own file scanning
Started after a driver or firmware update That update, straightforwardly

Use a copy tool that survives a bad link

Explorer is a poor instrument for a long transfer over a link you do not trust. It gives up on the first failure, discards what it had, and tells you nothing about which file failed or why. A retrying copy tool changes the failure from an opaque dialog into a log line you can act on, and it resumes rather than starting again.

  • Use robocopy for anything measured in gigabytes or hours, and keep its log.
  • It will not fix an unstable link, but it survives one, and its log tells you whether the failures cluster on one file or scatter across the set.
  • Failures scattered across many files point at the link. Failures on one file every time point at that file or the disk under it.
  • Run it from the machine with the better connection to the storage where you have a choice.

Before you blame the network

  • Check both disks. A failing sector on the source presents as a copy that stops at the same place every time, and no amount of adapter tuning will help.
  • Check whether a security product at either end scans every file as it is written, which adds latency under sustained load.
  • Check whether anything else is saturating the same link during the copy, including backups you did not schedule.
  • On a virtual machine, check the host as well as the guest. Offload and power settings exist at both levels.
  • Test in both directions. Offload faults are frequently one-way, and testing only one direction can send you looking at the wrong machine.

Changing advanced settings on a network adapter resets the link, and updating its driver disconnects the machine briefly. Do not do either through a remote session on a machine you cannot reach another way, and change one setting at a time so you can undo the one that made things worse.

Every code this article covers

Code What it points at Source
0x80070079 Win32 121, ERROR_SEM_TIMEOUT: the semaphore timeout period has expired. The wait for a response ran out Microsoft Learn
0x80070036 Win32 54, ERROR_NETWORK_BUSY: the network is busy Microsoft Learn
0x80070037 Win32 55, ERROR_DEV_NOT_EXIST: the specified network resource or device is no longer available Microsoft Learn
0xC00000C9 Microsoft publishes no meaning for this status value. In practice it accompanies a session that ended rather than one that timed out, so look for a restart, a failover or a device dropping the connection not published by the vendor
0xC000009A Microsoft publishes no meaning for this status value. It accompanies failures where the far end could not service the request under load, so treat it as a prompt to look at the far end rather than the client not published by the vendor

Confirm the fix worked

  1. Copy a large file, several gigabytes if you have one, and confirm it completes.
  2. Repeat the copy in the other direction, since offload faults are often one-way.
  3. Check the adapter’s error and discard counters after the copy and confirm they have not climbed.
  4. Leave a mapped drive connected for a working day and confirm it does not drop.
  5. If you changed adapter settings, confirm the ones that were not the cause have been put back.

Questions people ask about this

Is there anything to buy that fixes this?

No licence does. Every step here is a setting, a driver or a cable. If the far end is genuinely too small for the workload, that is a hardware conversation and not a licensing one.

Should I use robocopy instead of Explorer?

For large or long transfers, yes. It retries, it resumes and it reports which file failed and why, which turns an opaque copy failure into something you can act on. It does not fix an unstable link, but it survives one better.

Could antivirus be the cause?

It can contribute at both ends, since scanning every file as it is written adds latency under sustained load. Test rather than assume, and narrow any exclusion to the specific transfer path if one turns out to be justified.

Why did this start after a Windows update?

Usually because the update installed a new network driver. Check the driver date and version, and go back to the manufacturer’s build if the timing matches.

What do 0xC00000C9 and 0xC000009A mean?

Microsoft does not publish a meaning for either. They are worth noting because they usually accompany an abrupt end rather than a timeout, which moves the investigation to the far end, but do not let anyone tell you what they mean without a source.

Related error codes

Was this article helpful?

Your feedback helps us improve our documentation.

Related articles

Free Fix Error 0x80070576: clock skew stops Kerberos authentication on the client Free Fix Duplicate name on the network 0x80070034 and a broken Winsock TCP stack Free Fix Remote Desktop cannot verify the identity: certificate error 0x80090325 Free Fix Network adapter Code 10: this device cannot start in Device Manager
โ† Back to Knowledge Base