On Thu, Aug 7, 2014 at 9:45 AM, Agnetta Kamugisha < kamugis...@nccommunitycolleges.edu> wrote:
> Jeff, > > Check this link. > > We had to implement this workaround. > https://www.apachelounge.com/viewtopic.php?t=6037 FWIW, AH00344 after AH00356 shouldn't be interesting. The listening socket got closed as part of shutting down after the original problem. I hope to improve logging of that in the shutdown scenario so that people don't worry about that. This "netsh winsock reset" hint is interesting. With the report we have on 2.4.10, which tells us which handle goes bad (thus triggering AH00356), the bad handle is a rather generic handle. (CreateEvent(NULL, TRUE, FALSE, NULL);) When, or how often, do you need to use "netsh winsock reset"? Please confirm that you were seeing AH00356 without "netsh winsock reset". (I can imagine that there are situations with third-party winsock layers where "netsh winsock reset" helps with the AH00344 error which doesn't come right after the AH00356 error.) Thanks!
