httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Harrie Hazewinkel <harrie.hazewin...@jrc.it>
Subject fix for status bug
Date Tue, 01 Dec 1998 16:54:40 GMT
HI all,

Attached I have a patch that the status bug could fix.
I say could, since it makes use of my first proposed change
of an INTEGER in the scoreboard to recognise a vhost.

IMHO,
The disadvantages are:
--	NO use of the ptr towards the server_rec.
--	An internal mapping from the server_rec towards
	an integer is required.
--	At the request of the server-status you need
	to convert the INTEGER back into the server_hostname.

The advantages are:
--	Every process is not dependant on the 'shared memory'
	in which the server records are. So no bogus ptr's.
--	Even the request information of the old processes
	can be printed. Even when they are in "Graceful
	starting". If it would be fixed with checking
	on the child_status this information is not available!!
--	If a process cannot find the appropriate mapping
	it does not print anything.
-- (SIDE EFFECT) A numbering of the vhosts is also required for
	the SNMP module.

I tested it with multiple/simaltaneous "server-status?refresh"
requests and doing randomly "restarts and graceful restarts"



0- Harrie Hazewinkel --------------------------------------0
 mailto:harrie.hazewinkel@jrc.it       phone:+39+0332786322
 http://operations.ceo.org/~harrie/      fax:+39+0332785641
 postal: JRC of the E.C.  -       -  Ispra 21020 (VA) Italy
0----------------------------------------------------------0
Mime
View raw message