[prev in list] [next in list] [prev in thread] [next in thread]
List: net-snmp-bugs
Subject: [ net-snmp-Bugs-777013 ] SNMP loops over lseek(); read()
From: "SourceForge.net" <noreply () sourceforge ! net>
Date: 2004-02-25 22:33:10
Message-ID: E1Aw7aY-0003NI-00 () sc8-sf-web3 ! sourceforge ! net
[Download RAW message or body]
Bugs item #777013, was opened at 2003-07-24 11:48
Message generated for change (Comment added) made by slif
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=112694&aid=777013&group_id=12694
Category: agent
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Nobody/Anonymous (nobody)
Assigned to: Nobody/Anonymous (nobody)
Summary: SNMP loops over lseek(); read()
Initial Comment:
While snmpwalking or snmpgetting variables the snmp
daemon will suddenly loop thorugh
lseek(13,0xc9cd5080,0) =
-909291392 (0xc9cd5080)
read(0xd,0xbfbfe5a8,0x8) = 8 (0x8)
The position of the seek is incremented eight bytes
each time (note the read length). This happens over and
over again. The only code that I found might do this is in
agent/mibgroup/ucd-snmp/proc.c
however adding debugging code (using snmp_log())
doesn't cause anything to show up in the logs.
I do not have any proc directives in my snmpd.conf,
therefore I don't know why sh_count_procs would
suddenly get called.
When this happens the agent stops replying to requests.
Eventually the normal operating returns. I don't have a
good idea on the number of lines (I don't like crashing
production systems by killing disk space that fast) but
it's in the range of hundreds to thousands of loops
each time it does it. The system does typically run
1400-1500 processes, so..
One run starts
lseek(5,0xc0436984,0) =
-1069323900 (0xc0436984)
read(0x5,0xbfbfe84c,0x4) = 4 (0x4)
lseek(5,0xc047cea0,0) =
-1069035872 (0xc047cea0)
read(0x5,0xbfbfe848,0x4) = 4 (0x4)
lseek(5,0xc0436980,0) =
-1069323904 (0xc0436980)
read(0x5,0x80e3080,0x50) = 80
(0x50)
lseek(12,0xc0436980,0) =
-1069323904 (0xc0436980)
read(0xc,0xbfbfe7a0,0x4) = 4 (0x4)
lseek(12,0xc046da20,0) =
-1069098464 (0xc046da20)
read(0xc,0xbfbfe7a8,0x14) = 20
(0x14)
lseek(12,0xc047ce90,0) =
-1069035888 (0xc047ce90)
read(0xc,0xbfbfe744,0x4) = 4 (0x4)
lseek(12,0xc98ba2a0,0) =
-913595744 (0xc98ba2a0)
read(0xc,0xbfbfe748,0x18) = 24
(0x18)
lseek(12,0xc9c81000,0) =
-909635584 (0xc9c81000)
read(0xc,0xbfbfe6e8,0x8) = 8 (0x8)
lseek(12,0xc9c81008,0) =
-909635576 (0xc9c81008)
read(0xc,0xbfbfe688,0x8) = 8 (0x8)
and ends
lseek(12,0xc9d09878,0) =
-909076360 (0xc9d09878)
read(0xc,0xbfbfe568,0x8) = 8 (0x8)
lseek(12,0xc9d09880,0) =
-909076352 (0xc9d09880)
read(0xc,0xbfbfe568,0x8) = 8 (0x8)
partial lsof output shows
snmpd 84473 root 0u VCHR 5,0 0t6122595
8396 /dev/ttyp0
snmpd 84473 root 1u VCHR 5,0 0t6122595
8396 /dev/ttyp0
snmpd 84473 root 2u VCHR 5,0 0t6122595
8396 /dev/ttyp0
snmpd 84473 root 3w VREG 13,131084 154
31257 /var/log/snmpd.log
snmpd 84473 root 4r VCHR 2,0 0t0
7925 /dev/mem
snmpd 84473 root 5r VCHR 2,1 0xc045e50c
7924 /dev/kmem
snmpd 84473 root 6u PIPE 0xef393400 16384
->0xef394f80
snmpd 84473 root 7u PIPE 0xef394f80 16384
->0xef393400
snmpd 84473 root 8u PIPE 0xf1b444e0 16384
->0xef394080
snmpd 84473 root 9u PIPE 0xef394080 16384
->0xf1b444e0
snmpd 84473 root 10u IPv4 0xeaf8ad40 0t0
UDP *:snmp
snmpd 84473 root 11r VCHR 2,0 0t0
7925 /dev/mem
snmpd 84473 root 12r VCHR 2,1 0xc9d09888
7924 /dev/kmem
snmpd 84473 root 13r VREG 13,131084 32768
54672 /var/run
so 5 and 12 are both /dev/kmem
--
Michael Conlen
mconlen@neutelligent.com
----------------------------------------------------------------------
>Comment By: Michael J. Slifcak (slif)
Date: 2004-02-25 17:33
Message:
Logged In: YES
user_id=88697
please identify the scalar or table that snmpget will
trigger this behavior.
----------------------------------------------------------------------
Comment By: Nobody/Anonymous (nobody)
Date: 2003-07-24 11:50
Message:
Logged In: NO
I should have noted,
FreeBSD 4.8-RELEASE on i386 and ucd-snmp 4.2.1 and net-snmp
5.0.7
--
Michael Conlen
mconlen@neutelligent.com
----------------------------------------------------------------------
You can respond by visiting:
https://sourceforge.net/tracker/?func=detail&atid=112694&aid=777013&group_id=12694
-------------------------------------------------------
SF.Net is sponsored by: Speed Start Your Linux Apps Now.
Build and deploy apps & Web services for Linux with
a free DVD software kit from IBM. Click Now!
http://ads.osdn.com/?ad_id=1356&alloc_id=3438&op=click
_______________________________________________
Net-snmp-bugs mailing list
Net-snmp-bugs@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/net-snmp-bugs
[prev in list] [next in list] [prev in thread] [next in thread]
Configure |
About |
News |
Add a list |
Sponsored by KoreLogic