HOW-TO

HP-UX 101

sam
Text User Interface (TUI) for sys config (GUI version avail if DISPLAY set to working X server).				
in /usr/sbin
SAM is your friend.  It should be the first place to try if you are admin from another OS :)
Not as good as aix smitty, but better than solaris that have nothing! 
Use TAB key to move around, space key for select.  Arrow key works.

--0--

some of the major files location:
/etc/hosts
/etc/nsswitch.conf
/etc/fstab
/etc/lvmconf/vgNN   (binary files, but see all volume grp here)
/sbin/rc.config.d/
/sbin/rc.d/

System Setup


***** Initial system install checklist:   *****


Ensure that / and /usr are fairly big, as these are not not expandable 
unless in single user mode (or maybe onlineJFS).
Most other volumes still need to be unmounted before fs can grow, 
so estimating correct size is pretty important.
OnlineJFS should allow fs expansion w/o unmount like AIX, but more money.




Get latest of all 4 of BUNDLE11i, HWEnable11i, GoldApp and GoldBase and install them. 
Factory OS install included the latest release of all the above 4 patch bundles.

After bundles, java patches may still be needed.  See:
http://www.hp.com/products1/unix/java/patches/index.html 

Upgrading OS.
Default CD media allow for both initial setup and upgrade.
Upgrade has special procedure to install lot of depot first when system still 
running.  riptide was upgraded from 11.00 to 11.11, seems to run fine.


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

Update HP-UX from 10x and 11.0 to 11i:
install update-ux software from CD, then run 
/usr/sbin/update-ux -s /cdrom HPUX11i-OE.

Refer to Install and update guide in "Instant Documentation" CD.

--
HP-UX versions
PA-RISC: 10.20 (EOL), 11.00 (11), 11.11 (11i),               11.23 (ca 2005/05)
Itanium:                                       11.20, 11.22, 11.23


Boot Time Troubleshooting

Recovery of unbootable system. Default setup CD has option to do this. It may help when VxFS has crashed some critical OS partitions. booting HP-UX machine into single user mode. http://www.unixhub.com/docs/hpux/hpux_boot.html (Connect dumb terminal to serial console) Press and hold ESC at the beginning of the boot process (it should prompt, but if not, just hit ESC like crazy). In the PDC (firmware processor dependent code), enter BO to boot. (or BO P0 isl), where P0 is the boot device to use, whohwn by SEArch. When it ask whether to use interactive ISL (initial system load), say yes. a the ISL prompt, enter : hpux -is [ thre is a 'hpux -lm' for maintenance mode also it probably doesn't mount the / fs using normal method and thus VxFS stuff will not work -lm = lvm maint mode, sometime -is won't work when lvm is corrupted. -lq = lvm, ignore failed mirror drive allow booting of machine when one of the mirrored drive have failed. ll = list content of /stand, thus see filename of alternate kernel usable via eg -lm /stand/vmunix.alternate ] After this, should see these messages and the login prompt: Boot : disk(scsi.6;0)/stand/vmunix 966616+397312+409688 start 0x6c50 Kernel Startup Messages Omitted INIT: Overriding default level with level 's' INIT: SINGLE USER MODE WARNING: YOU ARE SUPERUSER!! # NOTE: The VGA monitor + pc keyboard may show some optinos. Hit F4 to toggle remote management on (with * on it). This makes serial console usable. When booted in single user mode with corrupted fs, vgchange can be used to tell Veritas Vol Mgrt which device to use. This is sometime needed even in non-mirrored disk.
vgchange -a y /dev/vg00		
lvlnboot -v /dev/vg00

fsck -F vxfs -o full,nolog /dev/vg00/rvol8
		# fsck for drives that have lot of errors, this method
		# may give more instructions to cmd of how to do the cleaning.
		# by specifying -F vxfs, allow use of intent log (journaling) 			
		# feature of JFS and thus completes much quicker.



************************************************************
clear out attention light in front panel

thru serial console session:

^B , enter 2x
get into gsp mode (Guardian Service Processor cli)
	give access to bios level control of the machine, 
	view vfp (virtual front panel), clear hw err log, etc.

sl	# sys log
e	# errors
n	# no filter
q	# quit
	
co	# exit back to console
************************************************************

---------- PDC commands ----------


PDC - Processor Dependent Code
Accessed at boot up time by the hardware to boostrap the 
system.  Step 1 of booting.
Only accessible during boot time.
Need to hit any key when prompted in a 10 seconds windows
during boot process to enter the PDC interaction session.
The window is after all hardware has been checked and before 
OS bootstrab process begins.

SEA	Search and display bootable device.
	Booting from CDROM may need to use this to find 
	device path to use for booting.

BO		boot, use default dev
BO PRI		boot, using primary (default) dev
BO SEC		boot, using secondary disk.  Simple to do dual boot.
BO 0/0/2/1.15	boot via specific path, see SEA.

IN	enter info menu
  ALL	display all info, there will be machine model strings 
	to determin what class of machine it is.
	eg  9000/800/A500-5X System Name: riptide
	is A-class.  some software like compiler bound to the 800 designation.
	Class info usually not shown in tag/model info in front os system :(

  ME	memory info, tell which bank has what kind of DIMM, vacancy.


---------- ISL commands ----------


ISL - Initial System Loader
Invoked by PDC so that it can load the kernel off /stand 
from the hard drive.

Interactions in here allow boot into single user mode, etc.
See HOWTO section and hpux command is in ISL.

hpux -is			# single user mode.  No password req!!
hpux -lm			# LVM maintenance boot mode.
hpux -lm /stand/vmunix.prev	# boot via specific kernel 
hpux ll				# list kernels avai in /stand

-0-
ISL> ?

    HELP          Help Facility
    LS            List ISL utilities
    AUTOBOOT      Set or clear autoboot flag in stable storage
    AUTOSEARCH    Set or clear autosearch flag in stable storage
    PRIMPATH      Modify primary boot path in stable storage
    ALTPATH       Modify alternate boot path in stable storage
    CONSPATH      Modify system console path in stable storage
    DISPLAY       Display boot and console paths in stable storage
    LSAUTOFL      List contents of autoboot file
    FASTSIZE      Sets or displays FASTSIZE
    800SUPPORT    Boots the s800 Support Kernel from the boot device
    700SUPPORT    Boot the s700 Support Kernel from the boot device
    READNVM       Displays contents of one word of NVM
    READSS        Displays contents of one word of stable storage
    LSBATCH       List contents of batch file
    BATCH         Execute commands in batch file
    LSEST         List contents of EST (Extended Self Test) file
    EST           Execute commands in EST (Extended Self Test) file

Enter 'LS' to see a list of the ISL utilities.

---------- GSP commands ----------


GSP - Guardian Service Processor.
BIOS level command, accessible even when machine is 
in Power Off state.  

Use serial console or LAN console, 
hit ^B
default is unsecured so just press enter for username and
password to gain GSP> prompt access.
Accessible all the time, it will not cause any interruption of running system.
If machine gets into boot process, user session to GSP will end.

HE = help, LI in there list commands:

==== GSP Help ============================================(Administrator)===
AC  : Alert display Configuration       PC  : Remote Power Control             
AR  : Automatic System Restart config.  PG  : PaGing parameter setup           
CA  : Configure asynch/serial ports     PS  : Power management module Status   
CL  : Console Log- view console history RS  : Reset System through RST signal  
CO  : COnsole- return to console mode   SDM : Set Display Mode (hex or text)   
CSP : Connect to remote Service Proc.   SE  : SEssion- log into the system     
DC  : Default Configuration             SL  : Show Logs (chassis code buffer)  
DI  : DIsconnect remote or LAN console  SO  : Security options & access control
EL  : Enable/disable LAN access         SS  : System Status of proc. modules   
ER  : Enable/disable Remote/modem       TC  : Reset via Transfer of Control    
HE  : Display HElp for menu or command  TE  : TEll- send a msg. to other users 
IT  : Inactivity Timeout settings       VFP : Virtual Front Panel display      
LC  : LAN configuration                 WHO : Display connected GSP users      
LS  : LAN Status                        XD  : Diagnostics and/or Reset of GSP 
MR  : Modem Reset                       XU  : Upgrade the GSP Firmware        
MS  : Modem Status                     

-0-

CO	return to console mode, exit GSP.  
	contrast to exit, which will logoff a telnet to the LAN console session.
LS, LC	Show and configure LAN console settings (IP, hostname, etc),
  	extremely useful, make sure to configure it!
TC	Send signal to OS, force it to do a core dump and reboot.
	Probably similar to Sun OBP sync cmd.
RS	Harshes restart of system, no OS interaction, no core.
	SUN OBM reset.
PC	allow turning machine on/off.  Can perform power cycle.
	GSP and LAN console will remain usable after power off.
	Contrast to SUN LOM and AIX serial access,
	just need to have power fed into machine.
PS	Monitor status of power supply, fan, etc.

SL	Show log, GSP level logging, with timestamp in UTC/GMT.
	Use this command to clear attention light.
CL	Show text message that has been displayed to the console.

VFP	display a virtual front panel, a mirror of the LED in front of machine.

Random Daily Tasks






changing hostname of a machine

/etc/rc.config.d/netconf	# top, see HOSTNAME="name"
/etc/hosts
/var/adm/inetd.sec		# may not be needed.

hostname NEWNAME (use this instead of reboot)

don't use set_parm
sam okay, but check files above.



move machine to new subnet, maybe use sam instead of config file changes for IP/NIC settings or else
need to edit multiple places.



creating /dev/zero

cd /dev
mknod zero c 3 0x000004
chown bin:sys zero
chmod 666 zero


Change system default timezone:
edit /etc/TIMEZONE to be:
TZ=PST8PDT
export TZ



Setting up NIS with predefinied list of servers.

/etc/rc.config.d/nameserver ::
NIS_DOMAIN-"titaniumleg.com"
(verify with domainname, immediate activation via domainname titaniumleg.com)

/var/yp/secureservers ::
10.215.10.255 10.215.10.17
10.215.20.255 10.215.20.21
10.215.30.255 10.215.30.6
10.215.40.255 10.215.40.199

/usr/lib/netsvc/yp/ypbind

Note that ypinit can only be used to setup master or slave server
not to init client config.







Storage

Disk

swap/paging 

swapinfo
swapon /dev/vg01/lvol2		# activate swap partition

/etc/fstab entry:
/dev/vg01/lvol2	/	swap	defaults 0 0 # swap device


/usr/sbin/sam goto 
Performance Monitors -> System Properties 
	there are tabs for cpu and memory info

Volume Management

creating root and swap volumes (using LVM)

[Adapted from Unix Sys Admin Handbook (USAH) p146.]
ioscan -fnC disk		# scan for disk dev path
pvcreate -B /dev/rdsk/c3t15d0	# mark disk for use as PV in LVM
dd if=/dev/zero of=/dev/rdsk/c3t15d0 bs=64k count=10
				# wipe out obsolete LVM marking on disk so that 
				# pvcreate will not complain, re run pvcreate cmd

mkdir /dev/vg01			
mknod /dev/vg01/group c 64 0xNN0000	# c 64 is disk major number
					# 0xNN0000 is uniq minor number hex, try 01, 02, ...
					# eg, chk against ls -l /dev/vg*/group

vgcreate -s8 /dev/vg01 /dev/dsk/c3t15d0	# create a new vol grp. w/ Phy Extend = 8 mb

mkboot /dev/rdsk/c3t15d0		# make it bootable disk

lvcreate -C y -r n -L 1024 /dev/vg01	# create contiguous dev lvol1 size 1024 MB
lvcreate -C y -r n -L 1024 /dev/vg01	# ditto, will be named lvol2 automatically

lvlnboot -r /dev/vg01/lvol1		# make lvol1 a boot partition
lvlnboot -s /dev/vg01/lvol2		# make lvol2 a swap dev


newfs -F vxfs /dev/vg01/rlvol1		# create the fs as veritas fs
swapon /dev/vg01/lvol2			# activate swap
vi /etc/fstab...

# swap addition like this req disk to have a boot partition, even on 2nd disk.
# sam may provide easier alternative.
# also, may need to set kernel param maxswapchunk which limit max swap size in byte
# calc as swchunk * DEV_BSIZE, whose def are 2048 and 1024, resp.



non "system" volumes can be created as, eg /usr/local:
lvcreate /dev/vg00	
lvextend -L 2000 /dev/vg00/lvol11 	# allocate 2 GB for the Log Vol.
newfs -F vxfs /dev/vg00/rlvol11		# lvol11 assigned dyn from above

Clean up procedure, reverse of create...
lvrmboot -r /dev/vg01			# remove root vol req for the LV.
					# undo lvlnboot -r and -s
lvremove /dev/vg01/lvol2		# remove the logic vol lvol2 from vg01
lvremove /dev/vg01/lvol1
vgremove vg				# remove the whole vol grp

bdf					# df, berkeley version

growfing the fs:
lvdisplay /dev/vg00/lvolX
# find out LE value
umount /fsmount
lvextend -l le_number /dev/vg00/lvolX	# where le_number is new LE value
					# or -L sizeInMB (final size)
extendfs  /dev/vg00/lvolX		# will fill up all allocated LE, JFS
fsadm -F vxfs size /fsmount		# increase fs for onlineJFS.
mount /fsmount

unless have onlineJFS, need unmount before extend.  /var, /usr need boot into 
single user mode (ISL hpux -is, -lm main mode won't work!)
lvextend, extendfs, then reboot to increase such fs.
Not sure how to do / !!
NOTE
/, /stand and swap require special policy of 'contiguous', and thus cannot be extended! (They usually correspond to lvol1 to lvol3). Wizar can do it by migrating to another disk and if onlineJFS is used.

Setup boot disk mirroring

Boot disk mirroring.

Adapted from HP-UX 11i system administration handbook and toolkit, 2nd Ed, by Poniatowski. p140
Here is the procedure to mirror root disk, which uses HPUX native LVM, and not the Veritas Volume Manager that the OS also ships with.
Warning
While this mirroring software is native to the HP-UX OS, it is an option that requires more money.
lvextend -m is the mirroring option that need license to be activated.
###
### ++ TODO: find out how to determine if machine has license for mirroring.
###


# eg, existing vg00 is in /dev/dsk/c1t2d0 (0/0/1/1.2.0)
# new drive for mirror is /dev/dsk/c2t2d0 (0/0/2/0.2.0)


The /etc/fstab before mirroring is:
/dev/vg00/lvol3   /        vxfs delaylog 0 1   
/dev/vg00/lvol1   /stand   hfs  defaults 0 1   
/dev/vg00/lvol4   /tmp     vxfs delaylog 0 2   
/dev/vg00/lvol5   /lhome   vxfs delaylog 0 2   
/dev/vg00/lvol6   /opt     vxfs delaylog 0 2   
/dev/vg00/lvol7   /usr     vxfs delaylog 0 2   
/dev/vg00/lvol8   /var     vxfs delaylog 0 2   
ioscan -funC disk
pvdisplay /dev/dsk/c1t2d0		# get lvm info of existing disk.
pvdisplay /dev/dsk/c2t2d0		# get err, no lvm def on it, raw disk.

pvcreate -B /dev/rdsk/c2t2d0		# add physical disk to be used by LVM
vgextend /dev/vg00 /dev/dsk/c2t2d0	# incorporate new disk to existing vg00

# do some checking
strings /etc/lvmtab			# see new disk used by lvm
lifls /dev/rdsk/c1t2d0			# see lot of boot util
lifls /dev/rdsk/c2t2d0			# no boot util, add next...

mkboot /dev/rdsk/c2t2d0			#
mkboot -a "hpux -lq" /dev/rdsk/c2t2d0	# SECONDARY disk, add AUTO, low quorum
mkboot -a "hpux -lq" /dev/rdsk/c1t2d0	# PRIMARY disk, low quorum (??)

lifls /dev/rdsk/c2t2d0			# see boot util, but not as extensive 
					# as primary disk.  It is ok.
					
###
###  The next steps will actually perform the mirror, 
###  and it will fail if -m mirroring option is not activated with valid license
###					

lvextend -m 1 /dev/vg00/lvol1 /dev/dsk/c2t2d0	# add mirror for lvol1=/stand
lvextend -m 1 /dev/vg00/lvol2 /dev/dsk/c2t2d0	# add mirror for lvol2=/ (root)
lvextend -m 1 /dev/vg00/lvol3 /dev/dsk/c2t2d0	# add mirror for lvol3=(swap)
# Repeat for all lvolX...
lvextend -m 1 /dev/vg00/lvol8 /dev/dsk/c2t2d0   # add mirror for lvol8=(/var)
lvextend -m 1 /dev/vg00/lvol7 /dev/dsk/c2t2d0   # add mirror for lvol7=(/usr)
lvextend -m 1 /dev/vg00/lvol6 /dev/dsk/c2t2d0   # add mirror for lvol6=(/opt)
lvextend -m 1 /dev/vg00/lvol5 /dev/dsk/c2t2d0   # add mirror for lvol5=(/lhome)
lvextend -m 1 /dev/vg00/lvol4 /dev/dsk/c2t2d0   # add mirror for lvol4=(/tmp)

setboot					# check current PDC bootpath
setboot -a 0/0/2/0.2.0			# set PDC> BO SEC (if needed)

###
### To undo the mirroring prep work:
###

vgreduce /dev/vg00 /dev/dsk/c2t2d0
pvremove /dev/rdsk/c2t2d0

###
###  Replacing a mirror disk  149
###

vgcfgrestore -n /dev/vgXX /dev/rdsk/c...                # vol grp conf restore
vgchange -a y /dev/vgXX                                 # change vol grp to avai
lable (-a y)
vgsync /dev/vgXX                                        # resync logical vol in
vol grp.

###
### Replace mirrored disk boot disk
### In addition to normal mirror disk replacements steps above, 
### add:
###

mkboot /dev/rdsk/c.t...
mkboot -a "hpux lq" /dev/rdsk/c...

reboot	# vgcfgbackup is run automatically on reboot

Other Volume Management Commands

HP-UX >= 11 default boot disk is under control of LVM - Logical Volume Manager.
Note
HP-UX shipws with TWO LVM. An native version for core OS partitions, and Veritas Volume Manager for use with data partitions. Filesystem created by LVM can be hfs (/stand, boot stuff) and vxfs (root, /var, /usr, etc). Further note that in addition to vxfs, HP talks about JFS and onlineJFS, see fsck.
To some extend, this is akin to Solaris, whereby some admin use the Sun Volume Manager (Disk Suite) for OS disks, and Veritas for data disks. HP-UX just happen to ship both pre-installed.
LVM  devices are /dev/vg00/lvolN
VxVm devices are /dev/vx/dsk/rootdg/volNN

LVM  cmds are lv... pv... vg...
VxVm cmds are vx...

---

LV = logical volume, similar to partition/slide of solaris.

ioscan -fnC disk		# scan for all hard drives (+ cd/dvd).

diskinfo

pvcreate -B /dev/rdsk/c3t15d0	# init a raw disk dev as dev for LVM -- ie make a PV.
pvremove /dev/rdsk/cXtXd0	# reverse/undo of pvcreate
pvdisplay /dev/dsk/c1t15d0	# display physical vol info (disk to vol grp info)
				# dev path from ioscan output

vgdisplay -v vg00		# display (verbose) vol grp info
				# vg00 = root vol grp (system files)
				# /etc/lvmconf list all vol grp in system

lvdisplay -v /dev/vg00/lvol8	# show logical to phy extents map for logical vol

# a few more commands. some sample eg in LVM mirroring section.

vgextend	# vol grp, add disk to
vgreduce	# vol grp, remove disk from
vgremove	# vol grp, remove it completely (all log vol in it will be gone)

lvcreate	# logi vol, create (initial setup)
lvextend	# logi vol, add space (physical extents) or mirror config
lvreduce	# logi vol, reduce space or remove mirror
lvchange	# logi vol, change allocation policy (congtiguous, etc)
lvremove	# logi vol, remove it completely (fs will be gone)
lvlnboot	# make lv bootable.

newfs, extendfs etc are applied to LV to make usable filesystem.



Filesystem


mounting cdrom:

mount -o ro -F cdfs /dev/dsk/c2t1d0 /mnt/cdrom

alternate approach (diff format? iona cd):
pfs_mountd &
pfsd &
pfs_mount /dev/rdsk/c2t1d0 /mnt/cdrom
pfs_umount /mnt/cdrom


Connectivity (Network)

NFS


automount typical auto startup (from ps):
/usr/lib/netsvc/fs/automount/automount -f /etc/auto_master -D ARCH=SC_HPUX	# old automount
/usr/lib/netsvc/fs/autofs/automountd -D ARCH=SC_HPUX				# newer autofs, /net may need some OS path in PARISC

/etc/rc.config.d/nfsconf				# set both to be sure!	
	AUTO_OPTIONS="-f $AUTO_MASTER -D ARCH=SC_HPUX"	#older automount
	AUTOMOUNTD_OPTIONS="-D ARCH=SC_HPUX"		#newer autofs
	see sambple config in hpux/nfsconf

/sbin/init.d/nfs.client	start 		# automount/autofs, can read the std out from this script

startup file is in /sbin/init.d/nfs.client

Note that if NIS map auto.master use undescore like auto_home, 
hpux won't be able to translate that to auto.home
As such, a slew of files need to be added in /etc/auto_home with conent of
+auto.home

/etc/auto.products		:: 	+auto.products
/etc/auto.automnt		::	+auto.automnt

Converting NIS server to have map with '.' should work.
But afraid break other OS platform, not well tested yet.
But Ohio NIS use that and it works fine!
Ass hole before created the SNAFU for me.

For compatibility issues, it is probably best to avoid both . and _ in the filename
There should not be reason why one can't use auto_master to have entries like:
/products 	autoproducts	# ie, no . or _

Not sure of the effect on auto_home and auto_direct

NFS

/etc/fstab

/dev/vg00/lvol8            /usr    vxfs   delaylog      0 2
netapp:/vol/vol1/users/sa  /mnt/sa nfs    rw,intr,hard  0 0


/etc/exports

/opt	-ro
/var	-ro

exportfs -a


NIC


lanscan  		# list all NIC, ifconfig -a doesn't work.
        -v		# verbose, show nic and driver, etc
ifconfig lan0		# almost all NIC are lanX
lanadmin -x 0		# tell what is current NIC config, eg duplex
				# 0 is NMID from lanscan.
lanadmin -X 100FD 0	# force settings, for this session only.
			
ioscan -fnC lan		# id card type and driver.
lsdev -C lan		# list all devices of Class lan

/etc/rc.config.d/hp*conf 	
eg 	hpbtlanconf	= btlan			# a-class build-in NIC
	hpgsc100conf	= btlan4 hsc j3515a	# some old box add on nic
	hpbasetconf	= btlan1
	hpbase100conf 	= btlan3 Internal (core) LAN card
	hppci100conf	= btlan5 "external" pci card A5230A
	hpsppci100conf	= btlan6 V-class pci lan card
	# Update this file to 
	# hard coded NIC settings (speed and duplex, eg 100 FD)
	# remember to add the card interface eg lan0 in the name 
	# section, or it will have no effect!
	# can also change settings via lanadmin -X 100fd 
	# but it will not live thru reboot.

Determining / setting nic duplex and speed
http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=684614


network driver patch for some old (a-class?) machine  btlan 
PHNE_22727	patch for btlan for hpux 11.11, older, stable, safe patch
PHNE_23465	latest btlan patch, have some warnings.  if work, autoneg will work better.

Network Config




change DNS servers
/etc/resolv.conf		# same as solaris/aix.
sam, networking and communications, DNS (BIND), DNS Resolver
use Menu, Actions, Set Default Domain...  to add "search" for other dns domains.  
config file edit is easier, only 1 file.

change ip/netmask:
sam, networking and communications, network interface cards
# this even allow for speed duplex settings!!  Advance has mac address and mtu.
/etc/rc.config.d/netconf 	# IP_ADDRESS[0]="10.10.10.123" SUBNET_MASK[0] 
				# BROADCAST_ADDRESS[0]="" (blank seems to work, auto calc?)
/etc/hosts 			# enter new IP of machine.

change routing:
sam, networking and communications, routes
/etc/rc.config.d/netconf	# ROUTE_DESTINATION[0]="default"
				# ROUTE_MASK[0]=""
				# ROUTE_GATEWAY[0]="10.10.10.1"	
				# ROUTE_COUNT[0]="1"		(aix consider it as 0)

Adding default router in command line:
route add default netmask 255.255.255.0 172.27.188.1 1


  

Serial/console port

Printing

To add a network printer, use sam,
printers and plotters, 
LP Spooler, 
Printers and Plotters, 

go to Actions menu, Add Remote Printer

Name of Queue to add                      --> Use printer name maching windows
Hostname of remote server                 -->  windows print server name
Name of queue on remote server            --> Windows printer name.
Type of print spooler on remote server    --> System V     (ie, don't check remote is BSD)


Sys Config

Package/Software management



swlist -l product 	
	list all product/patches in system
swlist | grep -e QPK -e GOLD
	determine if any Quality Pack or Gold Base patch has been installed 
	(service pack equiv).
	Apperently, all 4 of BUNDLE11i, HWEnable11i, GoldApp and GoldBase should be installed on the system.
	Factory OS install included the latest release of all the above 4 patch bundles.
swlist -l product | grep PH
	look for installed patches.  -l product seems to be needed for 11i 


swinstall -f mysoft -s /mnt/cd
	supposed to install software listed in file mysoft from cdrom

swinstall -s /home/sa/share/software-in-cd/hpux_aC++/aC++Compiler.800.depot
	needed to fully specify the depot file before the damn thing will recognize the file.  It is really just a tar file.

swinstall -s /export/tmp/bash.depot \*
	install everything in the depot (in this case, sure there is bash only).
	
	see email about where gnu tools are downloaded 
	[http://hpux.cs.utah.edu, depot format, some ins to /usr/local/bin, 
	other to /opt, into product specific subdirs!]


Patch


patch install eg:
(patch is obtained in a "depot" file format, which is really a tar.  Instruction presented in here is also in a text file inside the tar.)

http://www2.itrc.hp.com/service/patch/patchDetail.do?patchid=PHSS_20055&context=hpux:800:11:00

	1. Back up your system before installing a patch.

	2. Login as root.

	3. Copy the patch to the /tmp directory.

	4. Move to the /tmp directory and unshar the patch:

		cd /tmp
		sh PHSS_20055

	5. Run swinstall to install the patch:

[ to check that depot has been downloaded correctly:
  swlist -d @ /full/absolute/path/PHSS_20055.depot
]

		swinstall -x autoreboot=true -x patch_match_target=true \
			  -s /tmp/PHSS_20055.depot

	By default swinstall will archive the original software in
	/var/adm/sw/save/PHSS_20055.  If you do not wish to retain a
	copy of the original software, use the patch_save_files option:

		swinstall -x autoreboot=true -x patch_match_target=true \
			  -x patch_save_files=false -s /tmp/PHSS_20055.depot

	WARNING: If patch_save_files is false when a patch is installed,
		 the patch cannot be deinstalled.  Please be careful
		 when using this feature.

	For future reference, the contents of the PHSS_20055.text file is
	available in the product readme:

		swlist -l product -a readme -d @ /tmp/PHSS_20055.depot

	To put this patch on a magnetic tape and install from the
	tape drive, use the command:

		dd if=/tmp/PHSS_20055.depot of=/dev/rmt/0m bs=2k



Kernel/System Tuning


sysdef			# show many (but not all) kernel parameters
kmtune -l		# list all kernel parameters, -l for long list (in /usr/sbin)

ulimit -a	# display all limits, below are probably defaults.

core file size        (blocks, -c) 2097151
data seg size         (kbytes, -d) 1048576
file size             (blocks, -f) unlimited
max memory size       (kbytes, -m) unlimited
open files                    (-n) 2048
pipe size          (512 bytes, -p) 16
stack size            (kbytes, -s) 8192
cpu time             (seconds, -t) unlimited
max user processes            (-u) 76
virtual memory        (kbytes, -v) unlimited


ulimit -s 	# define stack size.
	governed by kernel static params and def vals
	maxsiz			         8388608 (8k)
	maxsiz_64bit		         8388608
	suggested val for small server 401604608 (max allowed, 383K)

ulimit -n 	# define max open files, def kernel values:
	maxfiles	2048	static		4096	# soft
	maxfiles_lim	2048	dynamic		8192	# hard



maxusers	# num of users, base param as multiplier by other params
		# does not actually govern logins.

npty		# pty for user login sessions.
nstrpty		# can set them to 1000
nstrtel

nproc		# max processes that can exist on system, 11i = static
		# def 20 + 8 * MAXUSERS (5988?)

use sam to define new params and build new kernel.
restart is needed.

Kernel parameters in default os is not very good for large server, For large java app servers, get HP tool Java2 Out-of-box and it will plug in all desireble parameters. Probably good for all server. http://www.hp.com/products1/unix/java/java2/outofbox/
download .depot, swinstal -s /full/path/to/joob_203_1100.depot -x autoreboot=true \*
Req reboot.  
/usr/sbin/kmtune will display the kernel params.  It should modify:
maxusers 512 
nproc 2048 
max_thread_proc 3000 
nkthread 6000 
nfile 3000 
maxfiles 2048 
maxfiles_lim 2048 
ncallout 6000 
maxdsiz 2063835136 
tcp_conn_request_max 2048

Adding International Language/Locale support

By default, all the basic support for international language are
installed.  Check with "locale -a".

The CDE language may need to be installed manually:
Insert CD2 of HP-UX 11i Foundation Operating Environment, then 
run SAM Software Install and add the CDE bundle of the desired language.

Hardware commands

model - give machine model info

Diagnostic tools

diagnostics tools:

netfmt -(lot of options) /var/adm/nettl.LOGnnn > /tmp/nettl.txt
	convert the network problem log from binary to text
	options may have been -l -N -n -f


stm, cstm xstm		
	hardware diagnostic program.  text menu, cli and X based.
	probably not well known, support people use only?

tusk
	some sort of network troubleshooting util.

Random Stuff

For those from Solaris background:
Env variables:
SHLIB_PATH			is used instead of Solaris LD_LIBRARY_PATH
LD_LIBRARY_PATH		for 64 bit version libs.  Solaris LD_LIBRARY_PATH64
LIB		There maybe something about the env var also.  Not sure.


dos2unix	
	Couldn't find such command tool, only found charconv.

charconv 
	convert input file to desired output format
	-u = unix  LF
	-d = dos   CRLF
	-m = mac   CR
	-i inputfile,  default is std in
	-o outputfile, default to std out
	--help more conversion info, including EBCDIC
	eg: cat dostext.txt | charconv -u > unixoutput.txt


Compiler:
	The bundled cc compiler in /usr/ccs/bin/cc_bundled only takes K&R function prototype syntax, not Ansi C.
	Moneyware aCC is Ansi C compliant and installs to /opt/aCC.


HP-UX Freeware


Info about installing hpux gnu freeware from hpux.cs.utah.edu. into NFS server central location.
No real need to use SAM to install the package and xfer to NFS server.
Can just do tar trick, see note below.  Warning: Some admin may see this as kludge as 
it bypass the software management tool.  Your choice :)

/import/hpux1100/usr/local$ cat dir.info

This dir is automounted as /import/hpux1100/usr/local
Various link would exist pointing to the parent dir of /import/hpux1100, such as /titaniumlegit.  
It could also be a symlink from slightly diff version of HP-UX 11.x


These files are gnu freeware depot files for HP-UX 11.00 from hpux.cs.utah.edu

The source is netapp:/vol/vol1/depot/hpux1100/usr/local  

The source depot files are hacked via untar and set of cp commands to put them into the desired 
usr/local/bin etc tree structure:

gunzip *.tar
mkdir untared
cd untared
for item in `ls ../*.depot`; do
        tar xf $item
done

mkdir -p cleaned/usr
cp -pR  */*/usr/* cleaned/usr

and finally the cleanred/usr/local dir is moved here.


---

the opt dir is extracted in similar manner:

mkdir -p cleaned/opt
cp -pR */*/opt/*/* cleaned/opt/

---

http://hpux.cs.utah.edu/
Software Porting and Archive Center for HP-UX
Freeware for HP-UX




[Doc URL: http://www.cs.fiu.edu/~tho01/psg/hpux.html]
(cc) Tin Ho. See main page for copyright info.


"LYS on the outside, LKS on the inside"
"AUHAUH on the outside, LAPPLAPP on the inside"