Bug 699571

Since its introduction on 22 Dec 2005, MonitorSelfAge on Linux has been 0.

src/condor_procapi/procapi.cpp's ProcAPI::getProcInfoRaw on Linux attempts to read proc in a loop, to absorb errors. Unfortunately, the sample_time (used to calculate age) is reset to 0 each time around the loop, including the first, and never properly re-set.

Fix is to move the setting of sample_time() into the loop, after the the procRaw is initialized (zero'd).

diff --git a/src/condor_procapi/procapi.cpp b/src/condor_procapi/procapi.cpp
index fcb28fa..bee8e36 100644
--- a/src/condor_procapi/procapi.cpp
+++ b/src/condor_procapi/procapi.cpp
@@ -492,12 +492,6 @@ ProcAPI::getProcInfoRaw( pid_t pid, procInfoRaw& procRaw, int &status )
                // assume success
        status = PROCAPI_OK;
 
-               // clear the memory of procRaw
-       initProcInfoRaw(procRaw);
-
-               // set the sample time
-       procRaw.sample_time = secsSinceEpoch();
-
        // read the entry a certain number of times since it appears that linux
        // often simply does something stupid while reading.
        sprintf( path, "/proc/%d/stat", pid );
@@ -508,7 +502,10 @@ ProcAPI::getProcInfoRaw( pid_t pid, procInfoRaw& procRaw, int &status )
 
                // in case I must restart, assume that everything is ok again...
                status = PROCAPI_OK;
+               // clear the memory of procRaw
                initProcInfoRaw(procRaw);
+               // set the sample time
+               procRaw.sample_time = secsSinceEpoch();
 
                if( (fp = safe_fopen_wrapper(path, "r")) == NULL ) {
                        if( errno == ENOENT ) {

Resolved upstream on V7_6-branch, 520a7f9c

    Technical note added. If any revisions are required, please edit the "Technical Notes" field
    accordingly. All revisions will be proofread by the Engineering Content Services team.
    
    New Contents:
C: Internal uptime stat reset to 0 before being reported
C: MonitorSelfAge is always 0
F: Internal uptime stat no longer reset to 0
R: MonitorSelfAge now reflects uptime

Note: Detect problematically with condor_status -long -master/-schedd/-any

Tested on:
$CondorVersion: 7.6.1 Apr 27 2011 BuildID: RH-7.6.1-0.4.el5 $
$CondorPlatform: I686-RedHat_5.6 $

$CondorVersion: 7.6.1 Apr 27 2011 BuildID: RH-7.6.1-0.4.el5 $
$CondorPlatform: X86_64-RedHat_5.6 $

$CondorVersion: 7.6.1 Apr 27 2011 BuildID: RH-7.6.1-0.4.el6 $
$CondorPlatform: I686-RedHat_6.0 $

$CondorVersion: 7.6.1 Apr 27 2011 BuildID: RH-7.6.1-0.4.el6 $
$CondorPlatform: X86_64-RedHat_6.0 $

# condor_status -long -master/-schedd/-any | grep MonitorSelfAge
MonitorSelfAge = 1684

available also using qpid-tool and cumin under slot(s)

>>> VERIFIED

(In reply to comment #7)
> available also using qpid-tool and cumin under slot(s)

under Overview->Performance

An advisory has been issued which should help the problem
described in this bug report. This report is therefore being
closed with a resolution of ERRATA. For more information
on therefore solution and/or where to find the updated files,
please follow the link below. You may reopen this bug report
if the solution does not work for you.

http://rhn.redhat.com/errata/RHEA-2011-0889.html