IcmpSendEcho2 error with errors with WSA_QOS_ADMISSION_FAILURE and ERROR_NOACCESS

I have an application that connects many servers. It works great for a few days, but will suddenly have many failures of one of two types:

WSA_QOS_ADMISSION_FAILURE (11010) "QoS error due to lack of resources"

or

ERROR_NOACCESS (998) "Invalid access to memory location."

It is strange that errors occur in beams. Those. all pings can fail in a few minutes with one of the above errors. Then it is cleaned. Later, all pins will not work for several minutes with a different error. It seems that they are not alternating.

This happens on Windows 2008 R2. I cannot reproduce it on my own, but if I wait a day or two, it will always happen again.

I checked and double-checked, then checked again to close all open descriptors.

This never happens when the application is first launched, so it does not seem to be related to finding or loading a DLL. And he corrects himself after a while, so it seems that he is not exhausted by the resource. And it has been working fine for several days, so it doesn't seem to be a problem using the API.

Into the abyss here. Does anyone have any ideas?

thank

+4
source share
1 answer

It turns out that the error code 11010is actually not WSA_QOS_ADMISSION_FAILUREfrom WinSock (which is not involved here), but a completely different value from the IP stack ICMP_ECHO_REPLYwith a much more significant value:

IP_REQ_TIMED_OUT (11010)

GetIpErrorString() " , FormatMessage, ".


, , 998.

NT Win32", , NT, ( 2005 ) 998 Win32 (ERROR_NOACCESS) :

STATUS_DATATYPE_MISALIGNMENT            ERROR_NOACCESS
STATUS_ACCESS_VIOLATION                 ERROR_NOACCESS
STATUS_DATATYPE_MISALIGNMENT_ERROR      ERROR_NOACCESS

, , - IOCTL ( ICMP-- , ), , , Win32.

, (, , , ), ICMP. , .

+8

Source: https://habr.com/ru/post/1538637/


All Articles