net-snmp

Hi,

today i compiled latest net-snmp 5.3.0.1 under cygwin which works fine,
but what i need is a way to get the uptime of the server using
'HOST-RESOURCES-MIB::hrSystemUptime.0' After reading some newsgroups and
the source code in 'snmplib/system.c' i saw that the #ifdef cygwin
return 0 because it is not implemented. Modern recent builds of Cygwin
(currently i am using Cygwin 1.5.19.4) has legacy support for
'/proc/uptime' which the get_uptime() function uses under linux.

Attached is a diff which only changes the #ifdef to linux || cygwin,
because cygwin supports '/proc/uptime' It works quiet well. Maybe it can
be integrated into the stable tree of net-snmp.

System:

mbroemme@...:~$ ssh 10.0.7.32 uname -a
CYGWIN_NT-5.2-WOW64 win-devel 1.5.19(0.150/4/2) 2006-01-18 11:15 i686
unknown unknown Cygwin

Without patch:

mbroemme@...:~$ snmpwalk -v 2c 10.0.7.32 | grep hrSystemUptime.0
HOST-RESOURCES-MIB::hrSystemUptime.0 = Timeticks: (0) 0:00:00.00

With patch:

mbroemme@...:~$ snmpwalk -v 2c 10.0.7.32 | grep hrSystemUptime.0
HOST-RESOURCES-MIB::hrSystemUptime.0 = Timeticks: (1158742) 3:13:07.42

--Maik

Re: Patch for hrSystemUptime and Cygwin. From: Dave Shield <D.Shield@cs...> - 2006-01-31 09:28 On Tue, 2006-01-31 at 02:31 +0100, Maik Broemme wrote: > Modern recent builds of Cygwin > (currently i am using Cygwin 1.5.19.4) has legacy support for > '/proc/uptime' which the get_uptime() function uses under linux. I'm a little concerned by that word "modern". How modern is modern? When was this feature introduced, and how widespread are earlier releases likely to be? More importantly, what happens if your patch is used on a system that does *not* include legacy support for /proc/uptime? Does it fail gracefully, or does it crash the agent? Dave

Re: Patch for hrSystemUptime and Cygwin. From: Maik Broemme <mbroemme@pl...> - 2006-01-31 09:38 Hi, Dave Shield <D.T.Shield@...> wrote: > On Tue, 2006-01-31 at 02:31 +0100, Maik Broemme wrote: > > Modern recent builds of Cygwin > > (currently i am using Cygwin 1.5.19.4) has legacy support for > > '/proc/uptime' which the get_uptime() function uses under linux. > > I'm a little concerned by that word "modern". How modern is > modern? When was this feature introduced, and how widespread > are earlier releases likely to be? > More importantly, what happens if your patch is used on a > system that does *not* include legacy support for /proc/uptime? > Does it fail gracefully, or does it crash the agent? > Well every Cygwin build i used since two years (before i did not use Cygwin) has support for '/proc/uptime' At this moment the patch does not check if '/proc/uptime' exist, sorry. :( I can implement this and return zero if it not exist. The '/proc/uptime' support under Cygwin is some kind of "virtual" which means that if the directory '/proc' not exist (it does not exist in a clean Cygwin installation) it is still accessible. > Dave > --Maik

Re: Patch for hrSystemUptime and Cygwin. From: Alex Burger <alex_b@us...> - 2006-01-31 16:46 Dave Shield wrote: > On Tue, 2006-01-31 at 02:31 +0100, Maik Broemme wrote: >> Modern recent builds of Cygwin >> (currently i am using Cygwin 1.5.19.4) has legacy support for >> '/proc/uptime' which the get_uptime() function uses under linux. > > I'm a little concerned by that word "modern". How modern is > modern? When was this feature introduced, and how widespread > are earlier releases likely to be? > More importantly, what happens if your patch is used on a > system that does *not* include legacy support for /proc/uptime? > Does it fail gracefully, or does it crash the agent? Looking at the Cygwin CVS, it looks like it has supported /proc/uptime for at least 3 years, 8 months. http://cygwin.com/cgi-bin/cvsweb.cgi/src/winsup/cygwin/ntdll.h?cvsroot=src Alex

Re: Patch for hrSystemUptime and Cygwin. From: Dave Shield <D.Shield@cs...> - 2006-01-31 16:57 On Tue, 2006-01-31 at 11:39 -0500, Alex Burger wrote: > Looking at the Cygwin CVS, it looks like it has supported > /proc/uptime for at least 3 years, 8 months. That sounds a reasonable lead time, and addresses my concerns. I've no further reservations about incorporating Maik's patch. Dave the Cautious

3 messages have been excluded from this view by a project administrator.