Description of Problem: After upgrading to last version (4.2.5) of ucd-snmp on several RH 7.3, 7.2 and 7.0 boxes (all running kernel 2.4.x) snmpd does not provide memory usage information. Version-Release number of selected component (if applicable): ucd-snmp-4.2.5-7.73.0 (RH 7.3, kernel 2.4.18-4 (from dist.)) ucd-snmp-4.2.5-7.72.0 (RH 7.2, kernel 2.4.17) ucd-snmp-4.2.5-7.70.0 (RH 7.0, kernel 2.4.17) Steps to Reproduce: 1. Install new ucd-snmp package 2. query (snmpwalk) enterprises.ucdavis.memory SNMP tree branch 3. compare results with /proc/meminfo Actual Results: $ snmpwalk placek mrtg enterprises.ucdavis.memory enterprises.ucdavis.memory.memIndex.0 = 0 enterprises.ucdavis.memory.memErrorName.0 = swap enterprises.ucdavis.memory.memTotalSwap.0 = -1 enterprises.ucdavis.memory.memAvailSwap.0 = -1 enterprises.ucdavis.memory.memTotalReal.0 = -1 enterprises.ucdavis.memory.memAvailReal.0 = -1 enterprises.ucdavis.memory.memTotalFree.0 = -2 enterprises.ucdavis.memory.memMinimumSwap.0 = 16000 enterprises.ucdavis.memory.memShared.0 = -1 enterprises.ucdavis.memory.memBuffer.0 = -1 enterprises.ucdavis.memory.memCached.0 = -1 enterprises.ucdavis.memory.memSwapError.0 = 1 enterprises.ucdavis.memory.memSwapErrorMsg.0 = Running out of swap space (-1) placek$ cat /proc/meminfo total: used: free: shared: buffers: cached: Mem: 1055617024 1028292608 27324416 0 193048576 601260032 Swap: 2154938368 11403264 2143535104 MemTotal: 1030876 kB MemFree: 26684 kB MemShared: 0 kB Buffers: 188524 kB Cached: 586660 kB SwapCached: 508 kB Active: 393048 kB Inact_dirty: 346076 kB Inact_clean: 50008 kB Inact_target: 157824 kB HighTotal: 130880 kB HighFree: 2208 kB LowTotal: 899996 kB LowFree: 24476 kB SwapTotal: 2104432 kB SwapFree: 2093296 kB Committed_AS: 44132 kB 8<=============================== Expected Results: snmpwalk should receive information as shown in /proc/meminfo
I can confirm this on several additional boxes running 7.2 and 7.3. The upgrade breaks memory usage. Ugh - first the disk directive blew up, then was fixed. Now the memory directive blew up....
L It looks like this is a bug in agent/mibgroup/ucd-snmp/memory.c I downgraded just that file to the one from 4.2.4 and things work. The patch that appears to have caused the problem was supposed to be a fix for values that were greater than a 32bit unsigned integer (either that or the fix to handle 2^32-1 that apparently was added to this patch a little later, judging from the CVS log): Its patch 548663. http://sourceforge.net/tracker/index.php?func=detail&aid=548663&group_id=12694&atid=312694
Hi, this is _the_ patch that worked for me -> # cat /usr/src/redhat/SOURCES/ucd-snmp-4.2.5-memory.patch --- ucd-snmp-4.2.5/agent/mibgroup/ucd-snmp/memory.c.orig Wed Jul 3 13:15:04 2002 +++ ucd-snmp-4.2.5/agent/mibgroup/ucd-snmp/memory.c Wed Jul 3 13:18:47 2002 @@ -272,9 +272,9 @@ for (i=0; i < MAX_ROW && *p; i++) { /* loop over rows */ while(*p && !isdigit(*p)) p++; /* skip chars until a digit */ for (j=0; j < MAX_COL && *p; j++) { /* scanf column-by-column */ - l = sscanf(p, "%u%n", &m, &k); + l = sscanf(p, "%lu%n", &m, &k); m /= 1024; - if (4294967295 > m) { + if (m > 4294967295) { *(row[i] + j) = 4294967295; } else { *(row[i] + j) = (unsigned) m; I reconstructed it from URL above, thanks guys, memory is showing fine now ... I added it into .spec file as Patch10 and rebuilded package. Spec diff _is_ also available on request. -- mozgy
Could we get a Rawhide SRPM/RPM with this patch? Please and Thank You. =)
*** Bug 67773 has been marked as a duplicate of this bug. ***
Errata will follow sometime next week with this patch. Read ya, Phil
*** Bug 69137 has been marked as a duplicate of this bug. ***
OK, took a couple of days longer, but a fixed ucd-snmp package is being built at the moment and the errata process has been initiated. The final errata should appear sometime the end of this week or around next (if all goes well ;-). Read ya, Phil
*** Bug 71385 has been marked as a duplicate of this bug. ***
Will this be released as errata soon?
*** Bug 73809 has been marked as a duplicate of this bug. ***
Is there any update in regards to this bug, besides the duplicate bug reports? There is a fix, but there has yet to be an Errata released.
I've bugged our QA department already, it's still in QA here (QA was swamped with work due to our latest Red Hat LInux 8.0 release). I won't make any promised anymore, but it's due to be released real soon now(tm). Sorry for the inconvenience and thanks for your patience. Read ya, Phil
*** Bug 77701 has been marked as a duplicate of this bug. ***
It has now been nearly six months since this bug was recognized and an errata was promised. According to RH policy, 7.3 is still a fully supported version and this is becoming somewhat ridiculous. When can we expect an errata? Can we, at the _very_ least, have an SRPM for this? I realize I can patch it myself, but this is not why I am paying RH for their Enterprise RHN service. Can someone please post an update? Thanks. Jim
I would like to 2nd that argument. What is going on? RedHat still supports the 7.3 release, which comes with ucd-snmp. If its too much of a burden for QA to verify the ucd-snmp patches, then at least backport the net-snmp program from 8.0. Either way, there is a supported OS that is not getting supported.
*** Bug 82249 has been marked as a duplicate of this bug. ***
I know this must sound like a lame excuse, but i'm still bugging our QA departement to finish up and release this errata. I'll see if i can force it to be released soon. In the meantime i'll provide the source and binary rpms on my people.redhat.com account: http://people.redhat.com/pknirsch/ Again, i can only apologize for this unexcusable delay, but as much as i'd like to i can't force the QA folks to work on it. Read ya, Phil
Did this pending errata ever get released for 7.3? It seems to be now two months since anyone complained at it not appearing... or are we supposed to have all moved to 8.0 by now?
*** Bug 89175 has been marked as a duplicate of this bug. ***
An errata 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 the 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/RHBA-2002-159.html
Talk about speedy resolution - 1 year and nearly three months after the bug was opened, you finally release an errata that claims to fix it. Off to verify those claims... (and you should've just waited until Jan 1, 2004, since 7.3 isn't supported anymore at that point - but thats another bone I have to pick...) This is getting as bad as the time cycle for big UNIX vendors to release patches.....I guess Red Hat has finally "arrived".