DHCP Server not able to handle the Traffic
Agarwal Vivek-RNGB36
RNGB36 at motorola.com
Sun Apr 26 12:17:29 UTC 2009
Hi
I ran the commands on the Server and I got the following results. Can
you please help me in pointing the root cause for packets getting
dropped
[root at ppdhcpsrv tmp]# vmstat
procs -----------memory---------- ---swap-- -----io---- --system--
-----cpu------
r b swpd free buff cache si so bi bo in cs us sy
id wa st
1 1 124 62768 233924 2731988 0 0 3 125 1 0 1 2
83 14 0
[root at ppdhcpsrv tmp]# iostat
Linux 2.6.18-53.el5 (ppdhcpsrv.menawimax.net) 04/26/2009
avg-cpu: %user %nice %system %iowait %steal %idle
1.33 0.00 1.85 13.63 0.00 83.19
Device: tps Blk_read/s Blk_wrtn/s Blk_read Blk_wrtn
sda 74.02 24.29 1002.99 258459635 10670953860
[root at ppdhcpsrv tmp]# netstat -s
Ip:
1708302478 total packets received
11 with invalid addresses
952 forwarded
0 incoming packets discarded
1708269698 incoming packets delivered
1795804172 requests sent out
33283 reassemblies required
16613 packets reassembled ok
526 fragments received ok
1052 fragments created
Icmp:
6736222 ICMP messages received
7 input ICMP message failed.
ICMP input histogram:
destination unreachable: 4169751
timeout in transit: 246989
echo requests: 2318982
echo replies: 500
4082493 ICMP messages sent
0 ICMP messages failed
ICMP output histogram:
destination unreachable: 1763511
echo replies: 2318982
Tcp:
624 active connections openings
696009 passive connection openings
831 failed connection attempts
154 connection resets received
2 connections established
132212788 segments received
70887573 segments send out
13902 segments retransmited
0 bad segments received.
4804 resets sent
Udp:
1560813436 packets received
1762616 packets to unknown port received.
6311155 packet receive errors
1718724673 packets sent
TcpExt:
87 invalid SYN cookies received
563 resets received for embryonic SYN_RECV sockets
1 packets pruned from receive queue because of socket buffer overrun
70 ICMP packets dropped because they were out-of-window
129838 TCP sockets finished time wait in fast timer
2 packets rejects in established connections because of timestamp
118986 delayed acks sent
17945 delayed acks further delayed because of locked socket
Quick ack mode was activated 257 times
8 times the listen queue of a socket overflowed
8 SYNs to LISTEN sockets ignored
11300 packets directly queued to recvmsg prequeue.
3425 packets directly received from backlog
66888 packets directly received from prequeue
275171 packets header predicted
1710 packets header predicted and directly queued to user
121225308 acknowledgments not containing data received
8742277 predicted acknowledgments
4501 times recovered from packet loss due to SACK data
Detected reordering 1 times using FACK
Detected reordering 1 times using SACK
Detected reordering 71 times using time stamp
57 congestion windows fully recovered
855 congestion windows partially recovered using Hoe heuristic
198 congestion windows recovered after partial ack
1568 TCP data loss events
44 timeouts after SACK recovery
2 timeouts in loss state
6255 fast retransmits
2828 forward retransmits
372 retransmits in slow start
1492 other TCP timeouts
77 sack retransmits failed
47 packets collapsed in receive queue due to low socket buffer
272 DSACKs sent for old packets
662 DSACKs received
9 connections reset due to early user close
584 connections aborted due to timeout
Regards
Vivek Aggarwal
+973-36583058
-----Original Message-----
From: sthaug at nethelp.no [mailto:sthaug at nethelp.no]
Sent: Sunday, April 26, 2009 11:57 AM
To: Agarwal Vivek-RNGB36
Cc: dhcp-users at lists.isc.org
Subject: Re: DHCP Server not able to handle the Traffic
> Iam using ISC-DHCP-3.0.5 installed on RHEL5
>
> When I take a tcpdump on the DHCP Interface; I saw around 1000 DHCP
Discovers at the interface but when I see /var/log/messages I can see
only around 500 DHCP DISCOVERs over there.
>
> Can anyone help me in debugging that where are the discovers being
lost and how can I check resolve that
Start by looking at system load. Check to see if the operating system is
losing UDP packets (which could happen), and pay particular attention to
the disk I/O operations. DHCP is often disk I/O bound, and an SSD or a
RAID controller with battery backed cache may help.
Steinar Haug, Nethelp consulting, sthaug at nethelp.no
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.isc.org/pipermail/dhcp-users/attachments/20090426/dd2ba83c/attachment.html>
More information about the dhcp-users
mailing list