4

I am getting SSL handshake failures randomly when number of concurrent users increases. Though the number is not that big, with only 100 users or so SSL handshake keeps failing randomly.

I have set the LogLevel to debug, but seemingly it doesn't tell me anything more significant apart from AH01998: Connection closed to child with abortive shutdown.

Usage of the memory is also not that high when the handshake failure happens. I am using apache V2.4.6 on CentOS 7, with mpm prefork.

Two things I would like to know:

  1. does the log entry for "abortive shutdown" has anything to do with SSL handshake failure?

  2. How can I debug it and understand what is causing the issue?

Chaminda Bandara
  • 2,067
  • 2
  • 28
  • 31
Vivek Vaghela
  • 41
  • 1
  • 1
  • 2

1 Answers1

6

The shutdown is the close of the SSL connection. It is thus a problem at the end and not the beginning of the connection and thus is not a SSL handshake failure. "Abortive" shutdown happens when the client simply closes the underlying TCP connection instead of doing an orderly SSL shutdown by sending SSL close notify messages. It can for example by caused by misbehaving clients (like robots) or by a crashed client.

Steffen Ullrich
  • 114,247
  • 10
  • 131
  • 172
  • Thanks for the insight! The clients are Anroid and iOS devices. I will try to debug the issue at client side, and see if there is anything that can be improved or fixed. – Vivek Vaghela Dec 06 '16 at 07:19