All about ntp

After a very grilling and fear-provoking experience with one of the consultants about the fact that ntp is not properly configured on few on our boxes and thats one of the primary reasons that nothing works in our environment :-). I had to get this right in my head...

Installing ntp on Linux/Solaris/AIX and OSX

Few facts about NTP:
Ntp is OS independent
NTP uses UTC as reference time
Even when a network connection is temporarily unavailable,
NTP can use measurements from the past to estimate current time and error

Stratum 0 clock - >> Reference Clocks -> Cesium Clock -> GPS
Stratum 1 clock - >> Top level NTP servers, directly connected to Stratum 0
Stratum 2 clock - >> Clients for Stratum 1
Stratum 3 clock - >> Clients for Stratum 2
---
---
Stratum 16 clock ->> Lowest level server


Peers: When servers synchronizes servers at same stratum server level, so they
may decide who has the higher quality of time and then can synchronise to the
most accurate, they are called peers.


NTP configuration model:
-Ntp can be configured in client-server model
-Peer to peer model,
-Also, a server may broadcast time to a broadcast or multicast IP addresses
and clients may be configured to synchronise to these broadcast time signals.


Few ntp commands:
#ntpq -p <-- show all peers used and configured together with their corner performance data.

bash-3.00# ntpq -p
remote refid st t when poll reach delay offset disp
==============================================================================
+pg913xs01.fe pg000xscrp01.fe 5 u 876 1024 377 0.31 -15.702 9.32
*pg913xs02.fe pg000xscrp02.fe 4 u 845 1024 377 0.23 5.291 4.26


Summary information includes the address of the remote peer,
the reference ID, the stratum of the remote peer,
the type of the peer (local, unicast, multicast or broadcast),
when the last packet was received, the polling interval, in seconds,
the reachability register,in octal, and the current estimated delay,
offset and dispersion of the peer, all in milliseconds.

#ntpdc
ntpdc> peers

#ntpdate -d 134.126.23.62 <--- Manually updating time with ntp server



Setting up and troubleshooting on AIX:

#1. Edit /etc/ntp.conf
#broadcastclient
server timeserver1
server timeserver2
server timeserver3
server timeserver4
driftfile /etc/ntp.drift
tracefile /etc/ntp.trace

#2. ntpdate 134.126.23.62 ( this is only required if you are way off )
9 Jul 21:27:48 ntpdate[299236]: step time server 11.16.4.62 offset -6059.104933

The offset must be less than 1000 seconds for xntpd to synch.
If the offset is greater than 1000 seconds,change the time manually on the client and run the ntpdate -d again.

#3. start xntpd
# startsrc -s xntpd
0513-059 The xntpd Subsystem has been started. Subsystem PID is 438386.

and

Edit uncomment the line in /etc/rc.tcpip
start /usr/sbin/xntpd -x "$src_running"

#4. Wait for atleast 6 mins before issuing, two lssrc results are listed below.
lssrc -ls xntpd


Look at the stratum value in two output listed below

bash-3.00# lssrc -ls xntpd
Program name: /usr/sbin/xntpd
Version: 3
Leap indicator: 00 (No leap second today.)
Sys peer: pg913xsfed02.ted.org
Sys stratum: 5 <------- this is good
Sys precision: -18
Debug/Tracing: DISABLED
Root distance: 0.152100
Root dispersion: 1.015091
Reference ID: 11.16.4.87
Reference time: ce014349.d0b4f000 Thu, Jul 9 2009 21:34:17.815
Broadcast delay: 0.003906 (sec)
Auth delay: 0.000122 (sec)
System flags: pll monitor filegen
System uptime: 279 (sec)
Clock stability: 0.000000 (sec)
Clock frequency: 0.000000 (sec)
Peer: time4.apple.com
flags: (configured)
stratum: 2, version: 3
our mode: client, his mode: server
Peer: pg913xsfed02.ted.org
flags: (configured)(sys peer)
stratum: 4, version: 3
our mode: client, his mode: server
Peer: pg913xsfed01.ted.org
flags: (configured)(sys peer)
stratum: 5, version: 3
our mode: client, his mode: server
Subsystem Group PID Status
xntpd tcpip 438386 active


bash-3.00# lssrc -ls xntpd
Program name: /usr/sbin/xntpd
Version: 3
Leap indicator: 11 (Leap indicator is insane.)
Sys peer: no peer, system is insane
Sys stratum: 16 <------- this is not good
Sys precision: -18
Debug/Tracing: DISABLED
Root distance: 0.000000
Root dispersion: 0.000000
Reference ID: no refid, system is insane
Reference time: no reftime, system is insane
Broadcast delay: 0.003906 (sec)
Auth delay: 0.000122 (sec)
System flags: pll monitor filegen
System uptime: 10 (sec)
Clock stability: 0.000000 (sec)
Clock frequency: 0.000000 (sec)
Peer: time4.apple.com
flags: (configured)
stratum: 16, version: 3
our mode: client, his mode: unspecified
Peer: pg913xsfed02.ted.org
flags: (configured)
stratum: 4, version: 3
our mode: client, his mode: server
Peer: pg913xsfed01.ted.org
flags: (configured)
stratum: 5, version: 3
our mode: client, his mode: server
Subsystem Group PID Status
xntpd tcpip 438386 active



Setting up on Linux:
#1. Edit /etc/ntpd.conf
server timehost1
server timehost2
server timehost3
server timehost4
driftfile /var/lib/ntp/drift

#2. /etc/init.d/ntpd start


Setting up on Solaris:
#1. Edit /etc/inet/ntp.conf
server timehost1
server timehost2
server timehost3
server timehost4
driftfile /var/lib/ntp/drift

#2. /etc/init.d/xntpd start
#3. svcadm refresh svc:/network/ntp



Setting up on OSX
#1. Edit /etc/ntp.conf
driftfile /var/lib/ntp/drift
server timehost1
server timehost2
server timehost3
server timehost4

#2. sudo /System/Library/StartupItems/NetworkTime/NetworkTime restart


---------------------------------------------------------------------------------------------------------


Problem: NTP daemon starts ok but dies after few minutes
Solutions:
1. Check the date on the machine. If it shows a strange date they could be missing /unix or /vmunix.
2. Check the TZ variable. Often a timezone variable on the client that is different than the server can cause this problem.
3. Make sure "broadcast client" line is commented out of /etc/ntp.conf.
4. How much is the time off? If it is >1000 seconds then NTP won't stay active. To correct this, run ntpdate serveripaddress.


Problem: No server suitable for synchronization found.
Solution:
If you start xntpd on a server and run ntpdate on a client to set the client's time with that of the server,
it will not update the client unless the xntpd daemon has been active for 6 minutes or longer.

Remove a file with a dash as first character or with special characters

Recently came across a file with dash in front and couldn't able to delete it with simple rm or rm with quotes.

Special characters files can be deleted using, if you know or can see all the special characters. For example pankaj%76$

rm "pankaj%76$"

However if there is a file like -pankaj%76$ (dash in front of it). Use
rm ./-pankaj%76$
or
rm -- -pankaj%76$

Using ./ or -- prevents the dash from coming in front of the filename and also being interpreted as rm command option.

How to identify the device(lun/scsi id) on Solaris presented from IBM DS4000?

How to identify the device(lun/scsi id) on Solaris presented from IBM DS4000?




Snapshot of DS4000:
Example: We are creating a LUN of 107GB and making it available only for hostname app1. The scsi id is 29





Array -> Logical Array - Drives


Solaris command:
root@dam-app1 # luxadm display /dev/rdsk/c4t600A0B80004715700000080F47FE4DEEd0s2
DEVICE PROPERTIES for disk: /dev/rdsk/c4t600A0B80004715700000080F47FE4DEEd0s2
Vendor: IBM
Product ID: 1815 FAStT
Revision: 0914
Serial Num: SP74850349
Unformatted capacity: 110160.000 MBytes
Write Cache: Enabled
Read Cache: Enabled
Minimum prefetch: 0x1
Maximum prefetch: 0x0
Device Type: Disk device
Path(s):

/dev/rdsk/c4t600A0B80004715700000080F47FE4DEEd0s2
/devices/scsi_vhci/ssd@g600a0b80004715700000080f47fe4dee:c,raw
Controller /devices/pci@0/pci@0/pci@8/pci@0/pci@2/SUNW,emlxs@0/fp@0,0
Device Address 201400a0b8471570,1d
Host controller port WWN 10000000c9722faa
Class primary
State ONLINE
Controller /devices/pci@0/pci@0/pci@8/pci@0/pci@8/SUNW,emlxs@0/fp@0,0
Device Address 202500a0b8471570,1d
Host controller port WWN 10000000c96307cc
Class secondary
State STANDBY


Device Address: 1d is in the hex decimal, converting it to decimal is 29. This is the same lun id which allotted to the host app1 from DS4000

Changing ftp default umask

Change ftp default umask

Step #1. Edit /etc/inetd.conf
from
ftp stream tcp6 nowait root /usr/sbin/ftpd ftpd
to
ftp stream tcp6 nowait root /usr/sbin/ftpd ftpd -u 007


Step #2. refresh -s inetd (AIX) or do a nohup on inetd