Vous êtes sur la page 1sur 17

Technical Tips and Papers

The information below is a collection of technical tips and tools for HP-UX in the form of questions and answers.All of the commands have been tested on the versions of HP-UX specified.Most of the commands are non-intrusive.That is to say, they make no permanent changes to your system.However, for those few commands below that do make permanent changes, be very careful.The author does not guarantee this document to be error-free.Please provide comments and/or questions to pete@bobcatconsulting.us. The questions included below were selected to illustrate some of the more obscure aspects of HP-UX.As such, the answers to those questions form a body of esoteric arcana heretofore solely within the province of gurus and wizards.Read on and join the club.If you would like to submit anything to this collection, or would like to see a particular question answered, please send it to pete@bobcatconsulting.us. This page was last updated on August 12, 2012. Configuration Information Im collecting system configuration information from all my systems and I would like to get it into a spreadsheet.What is the best way to do this? How big can a Version 2.x LVM volume group be? What is the System Management Homepage, and how do I access it? Im used to the pre-11.31 device file names and IO paths.How can I match what I remember with what I see on my 11.31 system? How do I tell which networked printer is which? How can I get more information about a file system that is already mounted on my system? What is my systems serial number? How much memory does my system have and how is it configured? What revision of sendmail is running in the system? Which kernel does my system have, 32-bit or 64-bit? How can I tell what patches are installed? What are the values of my kernel parameters? Whats in my routing table? What is the status of my SNA link? What is the status of the performance tools installed on my system? How can I get a list of the users defined on my system without printing the whole password file? How can I find the status of the Fiber Channel cards in my system? Restart Subsystems How do I restart the measurement tools? When the syslog daemon hangs up, can I restart it? How do I restart sendmail if it gets stuck? If the system clock falls too far behind the NTP servers time, how do I get it to sync up? Useful HP-UX Tools How do I extract a sub-string from a parameter in a script? What does it mean when I see rc=2 in the /var/adm/cron/log file? How do I get the ps command to work remotely using ssh? I want to reduce the size of a file system, but I remember bad experiences with this in the past.How can I accomplish this today? When I run my script remotely using ssh, it completes, then hangs.Is there any way to get around this, especially since I want to run this script on many systems? How do I comment out a line in a file, e.g. roots crontab, using a script? My syslog is far too big to edit (vi).Is there a way to get just todays entries? Can I know ahead of time if the installation of a patch set will require a reboot of the system? I had to replace one of the built-in disk drives connected to the internal Smart Array controller, and now the system cant access the new disk.What went wrong? Can I disable a swap device without rebooting my system? What will happen to my LVM environment when I update my 11.23 system to 11.31? I just updated my system to 11.31 and imported all my old volume groups.They appeared with the legacy DSFs.Why didnt I get the new persistent DSFs? I received my release media from HP and Im ready to do the update, but my systems are in another location.Can I do the update without transporting the physical media? I just received software from HP and it came in an ISO image file.The file name ends with .iso.What do I do with this? How can I copy my lp spooler environment from one system to another? Do I need OmniBack or TSM in order to use the DAT Autoloader on my system? How do I keep users off my system when I want to do maintenance? I have two systems with identical patch sets.Why does an HP-UX utility, such as sam, work OK on one system, but not the other? I just bought a new HP cell based server and the 2 cells came configured as 2 partitions.How can I make them into a single hard (nPar) partition? How can I make a depot file from an existing depot directory? I have an old fbackup tape.How can I tell when it was made? How do I wipe a disk clean of all information before removing it from the system? How can I expand a file system when I know I have more space available? Can I add a networked printer without going all the way into sam? How can I see or change any of the TCP timing parameters?

Generated with www.html-to-pdf.net

Page 1 / 17

Are there other network parameters that can be changed dynamically? Im adding a network interface and all I have is its IP address and broadcast address.What is its subnet mask? After my system crashes, is there any way to see what was on the console before the crash? How can I tell what is being dumped when my system crashes? Can I configure the crash dump? Can I regen my system without using sam? Can I get any more information from the ps command other than what the l or f options give? Should I inventory the media I get from HP? How can I nice a users whole session? How can I check the time on another system? How can I protect my system against a root drive failure? When my system is low on disk space, are there any likely places to look for files that can be safely removed? Is there any way to protect files in wide open directories, such as /tmp and /var/tmp? The bdf command hangs.Also, an ll / hangs too.What is going on? When bdf shows the root file system mounted to /dev/root, can I get rid of that? Little Known Facts Is there any online troubleshooting guide for HP-UX and its subsystems? Is there any other source for HP-UX tips other than this web site? What does the term password strength mean in terms of login security? Can I see the Epoch time without writing a C program? Once I have the Epoch time and I manipulate it, can I format it back into a readable string? What do all those numbers in a fiber channel hardware path mean? Whats all this talk I hear about new Daylight Savings time?Do I need to patch my system? I can email from my HP-UX system to my Outlook account, but when I set up roots mail to auto-forward to that same account, it doesnt go.What am I doing wrong? Why wont On-Line JFS work for me? What is waiting on inode? What are the ten most commonly exploited UNIX vulnerabilities? What are good, un-crackable passwords? What is the difference between the global run queue and the global priority queue? Are there any tricks to installing a new instance of HP-UX? How can I get a user to have the privileges of more than one group simultaneously? How can I search all the man pages for keywords? How can I synchronize the system clocks of all my systems? White Papers and Other Documentation And another thing a.k.a. Petes Soapbox

Configuration Information
Much of the information about HP-UX systems is transient, to the extent that it changes more frequently than documentation does.In order to maintain a current view of the systems issue the following commands at any time to retrieve the specified configuration information. Im collecting system configuration information from all my systems and I would like to get it into a spreadsheet.What is the best way to do this? One way would be to write a script to collect and parse the information you want, and deploy it to all your systems.When you gather all the output, assemble it into a comma delimited file, or CSV (Comma Separated Values).MS/Excel can easily import a CSV file and convert it to a spreadsheet.Be sure to name the created output file with a .csv extension when you copy it to the Windows system where Excel lives.Refer to the following web site for details on the CSV file format: http://creativyst.com/Doc/Articles/CSV/CSV01.htm
Back to top

How big can a Version 2.x LVM volume group be? The version 2.X volume groups are more dynamic than the version 1.0 volume groups.You can see a list of specifications by using /usr/sbin/lvmadm t.When creating a Version 2.X volume group, you only need two configuration parameters, the extent size and the maximum size of the volume group.In addition, you can preview the effect of the size parameters with the -E parameter in the vgcreate command: #/usr/sbin/vgcreate E V 2.0 s 4 This command will show the maximum size of a volume group with extent size of 4 MB. #/usr/sbin/vgcreate E V 2.0 S 128t This command will show the minimum extent size that can work with a volume group of 128 TB. #/usr/sbin/vgcreate V 2.0 s 4 S 128t vg09 This command will create a Version 2.0 volume group.
Back to top

What is the System Management Homepage, and how do I access it? The System Management Homepage, or SMH, is a web server resident on your HP-UX system that allows a system manager to view
Generated with www.html-to-pdf.net Page 2 / 17

and modify the systems configuration.From a web browser, connect to your system at port 2301, or port 2381 for secure access via https, and login as root.
Back to top

Im used to the pre-11.31 device file names and IO paths.How can I match what I remember with what I see on my 11.31 system? The ioscan command has a new parameter, -m <keyword>, which shows you the mapping between old and new naming conventions.When the keyword dsf is used, the mapping between legacy and persistent device special file names is displayed. When the keyword lun is used, the LUN to lunpath mapping is displayed.When the keyword hwpath is used, the LUN to lunpath to legacy hardware path mapping is displayed.
Back to top

How do I tell which networked printer is which? When you have networked printers, it is not obvious how each printer queue is associated with a device.The lpstat command does not help here.The problem with starting with jetadmin is that it keys off the IP name or address.If you want to start with the spooler queue name and figure out where it goes, you have to look somewhere else.There is no single file that lists this association.Each networked printer's IP name or address is hard-coded in its own model script. The following script gets all the info per each defined printer.Of course, this assumes that all queues are networked printers, not remote or local. lpstat -v | while read D F q device; do queue=${q%:} if [ $device = /dev/null ]; then IP=`grep ^PERIPH= /var/spool/lp/interface/$queue | cut -f2 -d=` echo "LP Queue name:"$queue echo "Destination:"$IP nslookup $IP | grep -E "Name|Address" echo /opt/hpnp/bin/addqueue i 1 h $IP -q $queue >> rebuild echo fi done
Back to top

How can I get more information about a file system that is already mounted on my system? To show the details contained in the superblock, use this command: #/usr/sbin/fstyp v <device-name> Where the device-name is the block special device file name which contains the mounted file system.The output will show the vxfs file system version, the block size and the fragment size among other less useful information. To find out if a file system is configured for large files, use this command: #/usr/sbin/fsadm F vxfs <mount-point> The output will display either largefiles or nolargefiles.
Back to top

What is my systems serial number? If you need to know your systems serial number, but cant get to the computer itself to look at the tag, try these two commands: #( cstm <<EOF select all infolog EOF ) | grep "System Serial Number" Or: #getconf MACHINE_SERIAL
Back to top

How much memory does my system have and how is it configured? Another use of the cstm utility is to get the memory configuration of your server.This command will report the total memory installed, and how it is configured: #echo selclass qualifier memory;info;wait;infolog | cstm
Back to top

What revision of sendmail is running in the system? The sendmail daemons version can be found by sending it the following obscure command: #echo '$=R' | /usr/lib/sendmail -bt -d0.4
Back to top

Which kernel does my system have, 32-bit or 64-bit? The HP-UX 11.x kernel can be either a 32-bit kernel or a 64-bit kernel.The following command has many options besides the one specified below. #/usr/bin/getconf KERNEL_BITS
Generated with www.html-to-pdf.net Page 3 / 17

The following example is undocumented and will return an integer indicating the PA-RISC Architecture version of your system. #/usr/bin/getconf SC_CPU_VERSION Match the output from the above command in the table below to identify your PA-RISC version. (0x20B) (0x210) (0x211) (0x214)
Back to top

523PA-RISC 528PA-RISC 529PA-RISC 532PA-RISC

1.0 1.1 1.2 2.0

How can I tell what patches are installed? A current patch list is useful in many troubleshooting scenarios.The following command can be issued on each system to generate such a list. #/usr/sbin/swlist l patch a patch_state \*.\*,c=patch Or, if you have HP-UX 11.0 and PHCO_27779 or HP-UX 11.11 and PHCO_27780, try this: #/usr/contrib/bin/show_patches
Back to top

What are the values of my kernel parameters? Use the following command to generate a list of common kernel parameter values.Alternately, sam can be used to generate a current list of all kernel parameter values. #/usr/sbin/sysdef Also, the following command can be used to generate a list of all kernel parameter values: #/usr/sbin/kmtune At HP-UX 11.23, the kmtune command is replaced by the kctune command.There is both a command line interface as well as a web interface.Refer to the White Papers and Other Documentation section below for a list of 11.11 and 11.23 defaults.
Back to top

Whats in my routing table? The following command will create a list of the entries in the current network routing table.The n parameter specifies that the entries should be displayed as numeric IP addresses.The use of this parameter saves the time in performing a name resolution on each address.This is particularly important in other forms of the netstat command that generate lots of output. #/bin/netstat -rn
Back to top

What is the status of my SNA link? Quite often, it is necessary to check the SNA status. The following command generates such status. The last command below generates the shortest summary. #/opt/sna/bin/snapadmin query_node #/opt/sna/bin/snapadmin query_pu #/opt/sna/bin/snapadmin d query_lu_pool
Back to top

What is the status of the performance tools installed on my system? The following command produces a summary of the performance tools status. #/opt/perf/bin/perfstat -v
Back to top

How can I get a list of the users defined on my system without printing the whole password file? The following table lists several commands for displaying user information. Command /usr/lbin/getprpw /usr/lbin/modprpw /usr/sbin/logins Description This command displays the user's protected password database settings in a trusted system. This command allows the superuser to modify a users protected password database settings in a trusted system. This command gives a formatted listing of the users defined in the password file in a non-trusted system.The format of the output can be controlled with command line arguments. This command lists users who are currently logged in to the system.It has many command line arguments to alter the content and display of the output. This command prints the current time, the length of time the system has been up, the number of users logged on to the system, and the average number of jobs in the run queue over the last 1, 5, and 15 minutes for the active processors.
Page 4 / 17

/usr/bin/who

/usr/bin/w

Generated with www.html-to-pdf.net

It is linked to uptime and prints the same output as uptime -w, displaying a summary of the current activity on the system. /usr/bin/listusers This command lists all users defined in the password file whose UID is greater than 100.It shows the user name and the GCOS field.Specific users or groups can be selected for display by command line arguments. /usr/sbin/ugweb
Back to top

This command starts the user and group account configuration tool.

How can I find the status of the Fiber Channel cards in my system? The following command will generate a summary of the status of a Fiber Channel interface. Such output is most useful when troubleshooting FC problems. #/usr/sbin/ioscan kfnC fc #fcmsutil <device name from above> #fcmsutil <device name from above> stat
Back to top

Restart Subsystems
Some of the HP-UX subsystems occasionally require a restart.Sometimes these processes are hung, or have simply stopped.The following list of commands identifies the most common subsystem restart mechanisms. How do I restart the measurement tools? Use the following command to restart the performance tools.This will stop the currently running tools and restart them. #/opt/perf/bin/mwa restart
Back to top

When the syslog daemon hangs up, can I restart it? Use the following script to restart the system logging daemon. #/sbin/init.d/syslog stop; /sbin/init.d/syslog start
Back to top

How do I restart sendmail if it gets stuck? Use the following script to restart the sendmail daemon. #/sbin/init.d/sendmail stop; /sbin/init.d/sendmail start
Back to top

If the system clock falls too far behind the NTP servers time, how do I get it to sync up? If the ntp functionality appears not to be operating, use the following command syntax: #/sbin/init.d/xntpd stop #ntpdate <server IP address> #/sbin/init.d/xntpd start
Back to top

Useful HP-UX Tools


The HP-UX system contains many useful tools, in the form of commands and run-time options for those commands.Documentation does exist for these tools.However, if a systems administrator is not aware of such tools, the documentation is not very helpful.The following list of questions and answers identifies those tools that will be most useful to the systems administrators. How do I extract a sub-string from a parameter in a script? Use the following syntax: #string=abcdefghijklmnopqrstuvwxyz #start=15 #length=5 #expr substr $string $start $length The output will be opqrs What does it mean when I see rc=2 in the /var/adm/cron/log file? This is an error (return code) generated by cron itself, indicating that it can not launch the command specified in the crontab entry. So far, I have found 3 such error codes: rc= 2 Error Description: Syntax error in the command or script

Generated with www.html-to-pdf.net

Page 5 / 17

126 127
Back to top

No execute permission Bad shell specification (#!/usr/bin/sh) or No such file

How do I get the ps command to work remotely using ssh? The ps command, like stty, who and others, is coded to write its output to a tty device.By default, the ssh client does not allocate a tty device.When you attempt to run ps on a remote system using ssh, you will get a message like this: ps: dont know which terminal to select However, the ssh t myhost ps command forces the ssh client to allocate a pseudo tty, allowing the ps command to operate.
Back to top

I want to reduce the size of a file system, but I remember bad experiences with this in the past.How can I accomplish this today? The Online JFS product has come a long way over the years.The current (11.23 and 11.31) file system reduction works as advertised. Assume that /dev/vg00/lvol9 is mounted to /var/adm/crash and is 4096 MB as reported by lvdisplay.Use the following command sequence to reduce the size of the logical volume and free 512 MB: #fsadm f vxfs b 3584m /var/adm/crash #lvreduce L 3584 /dev/vg00/lvol9 Note that you must shrink the file system BEFORE the logical volume.The size used in both commands must be an integer multiple of the physical extent size and is the new, smaller size: 4096-512=3584.
Back to top

When I run my script remotely using ssh, it completes, then hangs.Is there any way to get around this, especially since I want to run this script on many systems? Ive seen this behavior myself and, quite honestly, I do not yet understand why this happens. However, the following command sequence, at the end of your script, causes the script to terminate itself: #iam=`basename $0` #ps f | grep $iam | grep v grep | read uid pid ppid rest #echo killing $iam: $pid #kill $pid
Back to top

How do I comment out a line in a file, e.g. roots crontab, using a script? The problem is to pattern match just the line(s) you intend to comment out, and replace it with itself preceded by a #.Deleting the line is simpler, but not always desirable.Assume that you want to comment out a crontab entry that runs the log_rotate script: #crontab l | sed e s/.*log_rotate.*/#&/ > /tmp/mynewcrontab #crontab /tmp/mynewcrontab The trick here is the sed symbol & represents what is currently in the pattern buffer.
Back to top

My syslog is far too big to edit (vi).Is there a way to get just todays entries? Yes.Use the following command sequence to get all of todays syslog entries:
#SYSLOG=/var/adm/syslog/syslog.log #TODAY=`date #integer #integer

+%b"\ "%e`

N=`grep -n ^"$TODAY" $SYSLOG | head -n 1 | cut -d: -f1` N=$N-1

#cmd="1,$N"d #sed

"$cmd" $SYSLOG

For a quicker solution, and if you dont mind loosing the first entry for today, use this command sequence: #SYSLOG=/var/adm/syslog/syslog.log #TODAY=`date +%b"\ "%e` #cmd="1,/^$TODAY/"d #sed "$cmd" $SYSLOG Note that the sed command must be constructed as specified above, otherwise the symbols, $N and $TODAY will not evaluate correctly.
Back to top

Can I know ahead of time if the installation of a patch set will require a reboot of the system? Yes, you can.When you download a patch bundle from HP, its a pretty good bet that at least one of those patches will require a reboot. However, if you are researching a problem and find that you must download more than one individual patch, you may find yourself wondering what you just got.First, load all the individual patches into a single depot, either newly created or already existing.Then,

Generated with www.html-to-pdf.net

Page 6 / 17

query the depot for the reboot required status of all filesets contained therein: #swlist s <FQFN of depot> -l fileset a is_reboot This will show a column of fileset names contained within the depot and a column of true or false indicating if a reboot is required (true).
Back to top

I had to replace one of the built-in disk drives connected to the internal Smart Array controller, and now the system cant access the new disk.What went wrong? IfyouhavedisksmanagedbyaSmartArraycontroller,theSmartArraycontrollerassignsvirtualWWIDstoitsLUNs.Ifyouhaveadisk failure,thereplacementdiskwill(should)havethesameWWIDastheoriginaldisk.Ex:anSB40cstorageBladeslavedoffofan Integrity BL8x0c serverBlade. You wont have to run scsimgr, but you will have to run sautil to change the affected LUN from FAILED to OK. #sautil <device file of the SA controller> accept_media_xchg <LUN #> #Ex:sautil/dev/ciss1accept_media_xchg3 The 11iv3 LVM Admin manual shows a set of 2 commands that look like they do the same thing as the scsimgr command listed below.Ihavenotusedthistechnique,butitlooks like their 2 commands would do the same as the single command that I used. If the system creates another device file for the replacement disk, then the second command should clean up nicely. #scsimgr replace_wwid D <raw agile device file>#allows WWID replacement in this device file #Ex: scsimgr replace_wwid D /dev/rdisk/disk14 ThefollowingcommandreadsthenewWWIDfromnewlycreateddevicefileandputsitintooriginaldevicefile.Then,itremovesthe newdevicefile,ormaybesimplyrenamesthenewfilenametotheoriginalfilename. #io_redirect_dsf -d <original blocked agile device file> -n <new blocked agile device file> #Ex: io_redirect_dsf -d /dev/disk/disk14 -n /dev/disk/disk28 Thanks to RF for this info.
Back to top

Can I disable a swap device without rebooting my system? Prior to 11.31, you would have to remove the entry from the /etc/fstab file and reboot.As of HP-UX 11.31, there is a swapoff command available.If your system does not have this command (try man swapoff) you can download the SwapoffEnh bundle from the HP software depot at: software.hp.com and search for swapoff.Be sure to check the ITRC for the current vm cumulative patch for this command.As of 3/4/2010, it is PHKL_40130.
Back to top

What will happen to my LVM environment when I update my 11.23 system to 11.31? HP-UX 11.31 implements an Agile Naming Model for device special files, or DSFs.However, LVM will continue to support the legacy DSF names as well.Read the white paper titled LVM Migration from Legacy to Agile Naming Model HP-UX 11iv3 for an in depth explanation of the differences and benefits of the two naming models.
Back to top

I just updated my system to 11.31 and imported all my old volume groups.They appeared with the legacy DSFs.Why didnt I get the new persistent DSFs? There is a new option to the vgimport command in HP-UX 11.31, -N that will create the volume group using the persistent DSFs. You can either use vgexport and vgimport again, or use the /usr/contrib/bin/vgdsf utility to convert the volume group in place.Run the utility with no parameters to get its usage.
Back to top

I received my release media from HP and Im ready to do the update, but my systems are in another location.Can I do the update without transporting the physical media? Yes. There are several, Windows based tools that will copy the DVDs to an ISO image file. Such software is available both as freeware and as a purchased product.I used MagicISO available for download and purchase at http://www.magiciso.com. #Use your local Windows PC and your selected software to copy each release DVD to an ISO image file.Be sure to name the ISO files with names compatible with UNIX and easily identifiable. #Copy all ISO files to your HP-UX system.If you will be updating more than one system, it would be wise to copy the ISO files to an NFS server that is available to all systems to be updated. #Mount the first ISO image file.Refer to the question below for the details on how to do this. #Using the console port, run update-ux and specify the mount point as the source, eg. update-ux s /iso. #When prompted by update-ux, in another telnet session, mount the next ISO image file to the same mount point. Note that update-ux will, by default, update the OE that is currently installed.If you want to add additional software from the OE media you must specify the -i parameter and select it.If you want to add additional software from the application media, use swinstall using the ISO technique described above and below.
Back to top

I just received software from HP and it came in an ISO image file.The file name ends with .iso.What do I do with this? One option is to burn the .iso file to a DVD and install from that medium.However, there is another method if access to a DVD writer
Generated with www.html-to-pdf.net Page 7 / 17

and/or the system is problematic: #Create a logical volume in any volume group, slightly larger than the ISO image file. #dd if=<iso file> of=/dev/<vgname>/r<lvolname> bs=10240 #mkdir /iso #mount /dev/<vgname>/<lvolname> /iso #swinstall s /iso #unount /iso At this point, you can de-allocate the logical volume.Be aware that the depot may not be at the top level directory.Its best to look at the mount point and find the depot before executing the swinstall. There is a third method to access the ISO image if the system is HP-UX 11.31 with the ISOIMAGE-ENH bundle installed.Also, patches PHCO_39672, PHCO_39673, PHCO_39674 and PHCO_39675 should be installed.Verify with swlist.Once that is done, the ISO image can be mounted and used directly: #mkdir /iso #mount F cdfs <fully qualified iso file name> /iso #swinstall s /iso #umount /iso If the mount command generates an invalid operand error, try the following command to load the correct kernel module: #kcmodule fspd=loaded
Back to top

How can I copy my lp spooler environment from one system to another? The sam utility provides a facility to save and restore the entire lp configuration.It is saved in /var/sam/lp.To do this outside of sam, use the following command sequence.Assuming that you want to copy the lp configuration from sysa to sysb: #sysa: /usr/sam/lbin/lpmgr S #sysa: rcp r /var/sam/lp sysb:/var/sam/lp #sysb: /usr/sam/lbin/lpmgr R
Back to top

Do I need OmniBack or TSM in order to use the DAT Autoloader on my system? No. HP-UX 11.0 and 11i include a command to manipulate a media changer or autoloader. The command is mc and takes its parameters from the command line.That means that you can create a script to load the first tape in an autoloader and use tar, for example, to copy files to that tape, then change to the second tape and copy more files.However, because the basic tape utilities such as tar and cpio do not handle an end-of-tape condition gracefully you have to manage the amount of data transferred to each tape in your script. You will have to make sure that each invocation of a tar command does not exceed the capacity of the tapes in the autoloader.That being said, you might use the following command sequence to backup two large directories, where each directory is known not to exceed the capacity of the tapes in a six tape autoloader with all tapes in the magazine: #export LIBRARY=/dev/rac/c1t6d1# this is my autoloader device #mc s S1 d D1# move the tape from storage element 1 to data transfer device 1 #tar cvf /dev/rmt/1m /stuff1# copy files to tape #mc s D1 d S1# move the tape back to storage element 1 #mc s S2 d D1# move the tape from storage element 2 to data transfer device 1 #tar cvf /dev/rmt/1m /stuff2# copy files to tape #mc s D1 d S2# move the tape back to storage element 2 Note that in the above example, /dev/rmt/1m is the device name of the tape device and /dev/rac/c1t6d1 is the device name of the autoch device.The in-line comments indicate the purpose of each command.Refer to the mc man page for more command line parameters.These parameters allow you to check the status of the autoloader as well as determine the number of tapes currently loaded in the magazine.HP provides an installation and diagnostic tool known as the StorageWorks Library and Tape Tools.As of July09, the LTT is available for free download at: http://www.hp.com/support/tapetools.Look for the links identified as Documentation and Link to free download.
Back to top

How do I keep users off my system when I want to do maintenance? When it is time for system maintenance and you must keep users from logging on, HP-UX11.0 (with patches) and HP-UX 11i implement security procedures in the /etc/default/security file.This file should contain the line: NOLOGIN=1 to enable the functionality of the /etc/nologin file.When that file exists, new users are inhibited from logging on with a display of the contents of that file. When that file does not exist, no such restrictions are enforced. #banner System Maintenance in Progress > /etc/nologin
Back to top

Generated with www.html-to-pdf.net

Page 8 / 17

I have two systems with identical patch sets.Why does an HP-UX utility, such as sam, work OK on one system, but not the other? When patches are installed using the swinstall command, the installation process goes through several steps.Near the end of the process, patch filesets are configured, which is their final state.If the installation process is aborted, then patches can be one of several other states as well.Use the following command to check the state of all filesets, including patches, on your system: #swlist l fileset a state | grep v ^# | grep v ^$ | grep v configured The above command string generates a list of all filesets and their state, strips out all unnecessary lines and then strips out all configured filesets, leaving anything that may be installed, but not configured, corrupt or transient.Filesets that are installed should be configured using the swconfig command.Filesets in other states should be removed and re-installed.
Back to top

I just bought a new HP cell based server and the 2 cells came configured as 2 partitions.How can I make them into a single hard (nPar) partition? The system must have come along with sufficient hardware to have 2 nPars, such as a second core IO board and extra internal drives. However, if you want it all together in a single partition, you have to remove the 2nd partitions (Par 1) definition, and then add its resources to the 1st partition (Par 0).Lets assume that you have 2 cells, each connected to an IO chassis.Partition 0 contains chassis 1 and cell 1.Partition 1 contains chassis 0, cell 0.First, you have to make partition 1 inactive.Log on to that partition as root and issue: #shutdown R H 0 Note that the options are in upper case.Next, log on to partition 0 as root and issue the following commands: #parremove p 1 #parmodify p 0 a 0/0::y:ri #shutdown R 0 The first of these deletes the definition for partition 1.The second command adds Chassis 0, Cell 0 to partition 0.Finally, partition 0 is rebooted for reconfiguration.The parstatus command can be used at each step to verify the current partition configuration.
Back to top

How can I make a depot file from an existing depot directory? You can accumulate patches and software products into a single software depot using the swcopy command, allowing you to install those filesets in a single pass of swinstall.However, if you have to repeat that process, sometimes it is easier to package all those patches and/or software products into a single depot file. If your accumulated depot directory is /var/spool/sw, then use the following command to create stuff.depot in the /tmp directory: #cd /tmp #swpackage s /var/spool/sw x media_type=tape @ stuff.depot Of course, there must be sufficient space in the /tmp file system to accommodate the new file.
Back to top

I have an old fbackup tape.How can I tell when it was made? Have you ever had an old fbackup tape that you just cant identify?You can extract just the volume header with this command: #frecover -V /tmp/tapeheader -f /dev/rmt/<tape> Or, if you prefer the brut force method: #mt -f /dev/rmt/<norewind tape> fsf 1 #cat /dev/rmt/<tape> Note that <tape> will be something like 0m or 1m and <norewind tape> will be 0mn or 1mn.
Back to top

How do I wipe a disk clean of all information before removing it from the system? Quite often, system administrators find that they must replace a disk before it fails for upgrade purposes.When that disk contains company sensitive data, the disk should be wiped clean before removing it from the chassis.When that sensitive data is actually client information, there may be legal requirements to wipe the disk first, HIPAA for example.The following commands were extracted from the ITRC document 2200119109.The mknod command may not be necessary if the /dev/zero device already exists.Note that the block size specification in the dd command, bs=1024k, should be adjusted for the size disk that will be written to.Refer to the dd man page for further details.The disk wipe should only be performed after the disk has been successfully de-configured from LVM or any other volume manager. #mknod /dev/zero c 3 0x000003 #dd if=/dev/zero of=/dev/rdsk/<disk-to-wipe> bs=1024k
Back to top

How can I expand a file system when I know I have more space available? There are times when you must expand the size of a given file system.There are two steps in this process.First, extend the size of the logical volume using the lvextend command.Keep in mind that the size specified in this command is the new size, not the size increment.Second, you must extend the file system.If you have the luxury of un-mounting the file system, you can use the extendfs
Generated with www.html-to-pdf.net Page 9 / 17

command.However, if you have the OnLine/JFS product installed, you can use the fsadm command.Wile the extendfs command does not require that you specify the new size of the file system, the fsadm command does.It wants the new size specified in JFS blocks. The default value for the JFS block size is 1024. So, for example, when extending the /home file system, mounted from /dev/vg00/lvol5, to 500 MB use the following command sequence: #lvextend L 500 /dev/vg00/lvol5 #umount /home #extendfs F vxfs /dev/vg00/rlvol5 OR #lvextend L 500 /dev/vg00/lvol5 #fsadm b 512000 /home Note that the number of JFS blocks is the block size (default=1024) times the logical volume size in Mb. Alternately, the new size of the file system can be specified in MB as reported by the lvdisplay command: #fsadm b 500M /home
Back to top

Can I add a networked printer without going all the way into sam? Add a networked printer to the spooler from the command line: #/opt/hpnp/bin/multiaddqueue i 1 h <IPadr> -q <Qname> #/opt/hpnp/bin/addqueue i 1 h <IPadr> -q <Qname>
Back to top

How can I see or change any of the TCP timing parameters? The following command can be used to display the value of a particular TCP timing parameter.Other variations of this command will show all the parameters that can be modified by ndd, as well as set new values for these parameters. #ndd get /dev/tcp tcp_fin_wait_2_timeout
Back to top

Are there other network parameters that can be changed dynamically? The nettune command can be used to display and change the value of many network tunable parameters.The example below will generate a list of all available parameters, with their current, default, minimum and maximum values.Using the nettune command with no parameters generates a brief help screen. #/usr/contrib/bin/nettune -l
Back to top

Im adding a network interface and all I have is its IP address and broadcast address.What is its subnet mask? There are many good network calculators available on the web.The one I like is at: #http://www.subnetmask.info
Back to top

After my system crashes, is there any way to see what was on the console before the crash? Use the following command sequence to extract the system message buffer from a crash dump. #cd /var/adm/crash/crash.<sequence> #adb m vmunix . > msgbuf.out <<EOF msgbuf+8/s EOF
Back to top

How can I tell what is being dumped when my system crashes? The following command will display the current configuration of the crash dump facility.That is to say, it shows which elements of the HP-UX system will be dumped in case of a system crash, and how much disk space is required for such a dump.The same command can be used to re-configure the crash dump.This feature is part of HP-UX 11.0 and beyond. #/sbin/crashconf
Back to top

Can I configure the crash dump? Yes.The following file can be edited to configure the manner in which a crash-dump is saved to disk.Prior to 11.0, this file was called savecore.It had the same purpose, but a lot less functionality. #/etc/rc.config.d/savecrash
Back to top

Can I regen my system without using sam? In order to generate a kernel in preparation for the next reboot, edit the /stand/system file to include the necessary changes.Then, use the following command to set up the newly gen-ed kernel for installation at the next reboot.This technique is particularly useful when sam fails to provide this facility.This command is only available on HP-UX 11.0 and beyond.
Generated with www.html-to-pdf.net Page 10 / 17

#cd /stand; /sbin/config u system


Back to top

Can I get any more information from the ps command other than what the l or f options give? Yes.There are some interesting options to the ps command under XPG4 standards.For example, to get a listing of all processes with their virtual size, use the following command.A variation of this command is shown in the second example, as a way of showing the total virtual space demanded by the Oracle server.Refer to the ps man page for a complete listing of all XPG4 options. #export UNIX95=1 #ps e o pid,vsz,comm #ps u oracle o vsz | tail n +2 | awk { s += $1 } END { print s } The -H option formats the output in a process hierarchy format, showing parent-child relationships: #ps efH The -C optionallows you to specify the basename of a process: #ps fC <process_name> The -x option is not spe cific to the XPG4 feature set, but it does generate an extended command line: #ps fx The -ocomm,cpu,pcpu options show the cpu and percent cpu consumption for each running command: #ps eocomm,cpu,pcpu
Back to top

Should I inventory the media I get from HP? It is important to know the contents of all released software media.An index or table of contents can be built in a directory using flat ASCII files.Whenever you need to find a patch or software product, a simple grep across all files will identify the medium.Be sure to name the individual files appropriately. Use the date code, part number and/or sequence number.Bear in mind that the swlist command will list the entire contents of a secured CD medium regardless of your code-word permissions. #mkdir /TOC; cd /TOC #/usr/sbin/swlist s /cdrom > CD_0300_11.00_Appl_1of3
Back to top

How can I nice a users whole session? To nice a user, that is to say assign a nice value to his/her login shell, add "pri=" followed by a positive or negative number, as the first text in the gecos field of the /etc/passwd file. The nice value applies whenever the user logs in, but not for the su command.Commands such as chfn and finger are smart enough to ignore this text.
Back to top

How can I check the time on another system? Port 13 is the 'daytime' port on UNIX systems, as defined in the /etc/services file. You can check the time of a remote system by telnet-ing to port 13. This technique will NOT work on machines that have the 'daytime' service commented out of the /etc/inetd.conf file.Such machines might be configured that way as part of a security policy. #telnet that_host 13 Trying ... Connected to that_host. Escape character is '^]'. Wed May3 11:45:02 EDT 2000 Connection closed by foreign host. #telnet this_host 13 2> /dev/null | tail 1 Wed May3 11:45:20 EDT 2000
Back to top

How can I protect my system against a root drive failure? A physical backup of the root volume group is a good safety net in case of a total loss of the root drive and its mirror.As unlikely as this event may be, it is good to have a recent copy on hand for restoration.Use the following command to create an Ignite/UX backup.For very dynamic environments, it is wise to create a cron job for this command to run as frequently as is prudent for the number of changes expected on your server. #/opt/ignite/bin/make_recovery -A -v -i -C -d /dev/rmt/?mn 1>~/Recovery.log 2>&1 Where the options mean: -A Full root volume group -v Verbose output -iInteractive on booting from tape -C Create system status file -d Destination (always use the norewind device) HP recommends using a different utility:
Generated with www.html-to-pdf.net Page 11 / 17

#/opt/ignite/bin/make_tape_recovery xinc_entire=vg00 I v a /dev/rmt/?mn Where the options mean: -xinc_entire=vg00 [includes entire VG00] -a /dev/rmt/?mn[tape norewind device] -I[forces interactive recovery] -v[verbose messages]
Back to top

When my system is low on disk space, are there any likely places to look for files that can be safely removed? When in search of more disk space, there are two commands that identify and optionally remove extraneous file sets in the system.The freedisk utility is an interactive script that is used to remove non-dependant filesets that have not been accessed since their installation.The cleanup utility is used to remove old copies of code replaced by patches. #/usr/bin/cleanup #/opt/contrib/bin/freedisk v Refer to the man page for both commands before executing them.
Back to top

Is there any way to protect files in wide open directories, such as /tmp and /var/tmp? There is an interesting file protection mode called the sticky bit, which has a useful effect when applied to directories.If the sticky bit is set on a directory, files inside the directory may be renamed or removed only by the owner of the file, the owner of the directory, or the superuser even if the modes of the directory would otherwise allow such an operation.This is effective for temporary directories such as /tmp and /var/tmp which must be accessible to many authorized users. The following two commands have the same effect: #chmod a+rwxt /tmp #chmod 1777 /tmp allows anyone to create, read, and write files in /tmp, but only the file owner, the directory owner, or root can delete files.
Back to top

The bdf command hangs.Also, an ll / hangs too.What is going on? Sometimes, network based mount points, such as from NFS, Samba or CIFS, can go bad due to network issues (timeouts, noise, etc.). The HP-UX system does not know that the file server is no longer responding and there is still an entry in the /etc/mnttab file for the network mount.Any command that traverses the entire mount table will hang when trying to access the networked mount point.One way to get out of the hang problem is to rebuild the mount table. In the absence of a /etc/mnttab file, the mount command will reconstruct a new one.Of course, once this issue is resolved, the network problem that triggered this behavior must be resolved. #cd /etc #mv mnttab mnttab.orig #mount
Back to top

When bdf shows the root file system mounted to /dev/root, can I get rid of that? When doing certain troubleshooting activities, such as booting up in LVM maintenance mode, the root file system will be mounted t o /dev/root instead of /dev/vg00/lvol3. Sometimes that condition still exists after troubleshooting. Both a bdf and lvlvnboot v will show the bogus entry.One way to clear this up is to do the following: #cd /etc #mv mnttab mnttab.orig #mount a #bdf #lvlnboot -v
Back to top

Little Known Facts


The following paragraphs list some interesting observations and little know facts about HP-UX and Unix in general. Is there any online troubleshooting guide for HP-UX and its subsystems? Yes, HP maintains online assistance at this web site: http://www.hp.com/support/hpgt.Once there, select from the list on the left, or enter a product name in the search panel on the right, such as ignite.
Back to top

Is there any other source for HP-UX tips other than this web site? Yes, of course there are many web sites that have technical information.However, none are as nicely presented and useful as this one. Two other web sites that I have found to be particularly helpful are: http://blog.sourcedirect.com/category/unix/hpux and http://www.hpuxtips.es.
Back to top

Generated with www.html-to-pdf.net

Page 12 / 17

What does the term password strength mean in terms of login security? Password strength is a measure of the effectiveness of a password in resisting guessing and brute-force attacks. In its usual form, it estimates how many trials an attacker who does not have direct access to the password would need, on average, to guess it correctly. The strength of a password is a function of length, complexity, and unpredictability.A detailed explanation can be found on Wikipedia at: http://en.wikipedia.org/wiki/Password_strength.
Back to top

Can I see the Epoch time without writing a C program? Yes.Use the following command: date +%s.
Back to top

Once I have the Epoch time and I manipulate it, can I format it back into a readable string? Yes.You can use a lengthy algorithm to arithmetically convert seconds into days, months and years, but its much easier to let perl do the work for you: #now=`date +%s` #integer now=$now+86400# 1 days worth of seconds #perl -le 'print scalar(localtime('$now'))' Some of the time stamps in the trusted system database, such as the account expiration time, is stored as Epoch time.If you have a trusted system, use /usr/lbin/getprpw myuname to see what I mean.
Back to top

What do all those numbers in a fiber channel hardware path mean? Interpreting the hardware path in the ioscan output for Fibre Channel devices is different from interpreting ioscan output for other types of devices. The Fibre Channel hardware path format for a private loop device is as follows:

Each field in the hardware path is described below: Bus_Converter Adapter Protocol_Type Area Port One of several possible bus converters providing the interconnect to the bus to which the Fibre Channel adapter is connected. Module address of the Fibre Channel adapter. Fibre Channel protocol type. For Fibre Channel Mass Storage, the protocol type is FCP and is represented by a value of 8. The area portion (second byte) of the N_Port Identifier. For a private loop topology, the value is 0. When interpreting the hardware path for Fibre Channel Mass Storage devices, arrays are defined as having addressable controllers, and non-arrays are defined as not having addressable controllers. For hardware paths associated with non-array LUNs, the port element of the hardware path is set to the value 255 (255 for direct connect controller). For hardware paths associated with array LUNs, the port element of the hardware path is set to the Loop ID (for a private loop topology). For hardware paths associated with non-array LUNs, the bus element of the hardware path gets its value from the upper four bits of the port portion of the N-Port Identifier. For hardware paths associated with array LUNs, the bus element of the hardware path is set as: The bus number (0-7) for an array device using the logical unit addressing method The value of the upper seven bits of the 14-bit volume number for an array device using the volume set addressing method. For hardware paths associated with non-array LUNs, the target element of the hardware path gets value from the lower four bits of the port portion of the N_Port Identifier. For hardware paths associated with array LUNs, the target element of the hardware path is set as follows: The target number (0-15) for an array device using the logical unit addressing method. The value of bits 6-3 (0 is the least significant bit of the 14-bit volume number) for an array device using the volume set addressing method.

Bus

Target

Generated with www.html-to-pdf.net

Page 13 / 17

LUN

For hardware paths associated with non-array LUNs, the LUN element of the hardware path is the LUN ID (0-7) for the LUN. For hardware paths associated with array LUNs, the LUN element of the hardware path is set as follows: The LUN ID (0-7) of the LUN for an array device using the logical unit addressing method. The value of bits 2-0 of the 14-bit volume number for an array device using the volume set addressing method.

The following illustration is the Fibre Channel hardware path format for a public loop device:

Each field in the hardware path is described below: Adapter Domain Module address of the Fibre Channel adapter. Domain ID or Fabric ID (8 is reserved for private loop).A domain can be defined as two or more nodes connected by an interconnect subsystem. In a fabric environment, a domain is generally associated with the physical instance of a switch. The area portion (second byte) of the N_Port Identifier.In a fabric environment, the area ID is generally associated with a physical port on a switch. (For a private loop topology, the value is 0).In the above ioscan example, the area ID 19 is not the actual port number. Some switches place a fixed 1 in the upper four bits of the area field (0001xxxx) and shows the value of this field as hexadecimal. However, ioscan shows a decimal value for this field. Because of the fixed 1 in this position, you must subtract decimal 16 from the number 19 to get the actual port number, which is 3 (19 - 16 = port 3). Set to 255 for Peripheral Device Addressing. The bus element of the hardware path gets its value from the upper four bits of the AL_PA (not the Loop_ID). The target element of the hardware path gets its value from the lower four bits of the AL_PA (not the Loop_ID). In the above example, the AL_PA = 15.4 = E4 hex = 228 decimal = switch setting of 2). The LUN element of the hardware path is the LUN ID (0-7) of the LUN.

Area

Port Bus Target

LUN
Back to top

Whats all this talk I hear about new Daylight Savings time?Do I need to patch my system? Starting in 2007, U.S. Daylight Saving Time (DST) will begin on the second Sunday of March and end on the first Sunday in November. These changes to the schedule, the first in 20 years, will make U.S. Daylight Saving Time four weeks longer and could impact systems and processes. Daylight Saving Time is the practice of moving clocks forward one hour in the spring through summer to provide more daylight in the evening hours when outside activities are more common.The U.S. Energy Act of 2005 made this change to the DST policy in order to reduce energy consumption.Beginning in 2007, the clocks will move forward one hour on the second Sunday of March (March 11 in 2007) instead of the first Sunday in April.Come fall, they'll be set back an hour on the first Sunday of November (November 4 in 2007), rather than the last Sunday in October.The change will not affect Arizona (excluding the Navajo Nation), Hawaii, and all U.S. Territories, which do not observe Daylight Saving Time. In March of 2006, HP released the first set of DST patches.In February 2007, HP released a set of patches to accommodate the Mexico, Western Australia and Canada DST policies that supersede the previous patches.HP-UX 11iV3 will have this functionality included.The following table lists the DST patch IDs for the supported releases of HP-UX.These patches should be installed on all systems by 2007. Release Date Mar. 13, 2006 Mar. 29, 2006 Feb. 12, 2007 11.0 PHCO_34266 PHCO_34673 PHCO_35991 11.11 PHCO_34342 PHCO_34668 PHCO_35989 11.23 PHCO_34375 PHCO_34366 PHCO_35990

These patches can be installed while the system is up.There is no automatic reboot.Even though there are separate patches for the different supported releases of HP-UX, they all contain the same file: /usr/lib/tztab.The patches simply replace the tztab with an updated one.This /usr/lib/tztab is copied from the PHCO_35989 patch. For a brief explanation of how HP-UX keeps and reports time, see Time.

Generated with www.html-to-pdf.net

Page 14 / 17

Back to top

I can email from my HP-UX system to my Outlook account, but when I set up roots mail to auto-forward to that same account, it doesnt go.What am I doing wrong? First, you should check the /var/adm/syslog/mail.log file for any errors.Also, look in /var/mail/root to see if the mail is being returned to root on the server.Since the mail is delivered to your Outlook mail account when you simply send a message, there is probably nothing wrong with the /etc/mail/sendmail.cf file. However, the current implementation of sendmail will not autoforward if the forward file has world writeable permissions.Try this: #chmod 750 ~root/.forward
Back to top

Why wont On-Line JFS work for me? HP-UX 11i, as well as 11.0 with patches, requires a particular device file to enable On-Line JFS functionality.If this device does not exist, create it with: #mknod /dev/vxportal c 79 0
Back to top

What is waiting on inode? When multiple processes have the same file open, there are multiple file opens, but there is only 1 inode table entry.The file system allows only 1 process to have an IO active per inode/file.All other processes will show up as "waiting on inode".
Back to top

What are the ten most commonly exploited UNIX vulnerabilities? This information is provided by the Sans Institute: http://www.sans.org. What are the ten most commonly exploited UNIX vulnerabilities? #Poor system administration practices #Race conditions #Reusable/poor passwords #Flawed SUID programs (e.g., rdist, binmail) #HTTP server and CGI application vulnerabilities #Default "+" entries in the /etc/hosts.equiv file #NFS/NIS vulverabilities #sendmail program bugs #Buffer overruns (e.g., gets(), syslog()) #SUID shell scripts
Back to top

What are good, un-crackable passwords? Information from security audit companies has indicated that the best passwords to use, for security purposes, are passwords that utilize keyboard patterns to form meaningless words. Avoid the use of words that can be found in the dictionary and the use of digits to replace similar looking letters.For example, ru48ei39 is a simple keyboard pattern that will not be cracked, whereas f00tba11 will be cracked.
Back to top

What is the difference between the global run queue and the global priority queue? There are two performance metrics that are used to indicate a CPU bottleneck: the global priority queue (GBL_PRI_QUEUE) and the global run queue (GBL_RUN_QUEUE).Both metrics are captured by the MeasureWare agent and displayed by Glance/GPM.However, it is very important to understand the difference between these 2 metrics. The GBL_PRI_QUEUE is the average number of processes blocked on PRIORITY; that is, waiting on their priority to become high enough to get the CPU.This metric is calculated as the accumulated time that all processes spent blocked on PRIORITY divided by the interval time.Since this metric is calculated, it is possible for it to record uncharacteristically high values.Be sure to ignore spikes of 327.67 since this indicates an arithmetic error, not a very busy system. The GBL_RUN_QUEUE is the average number of "runnable" processes during the interval.This metric is derived from a kernel variable (avenrun) which is calculated by summing the number of runnable processes and averaging the samples over the last minute. Processes marked "runnable" include: #A process using the CPU at the time of the sample #A process waiting for the CPU at the time of the sample #A process paused on a "short disk wait" at the time of the sample Because this metric can include processes that are waiting for disk IO to complete, it is not a reliable CPU bottleneck indicator.Several standard Unix commands such as uptime, w and top display avenrun as the "1-minute Load Average.The GBL_RUN_QUEUE is normally a very small number.Larger than normal values for this metric can indicate CPU contention among processes.A high value for total global CPU utilization (GBL_CPU_TOTAL_UTIL) is another sign of possible CPU contention. It may be OK to have GBL_CPU_TOTAL_UTIL be near 100 percent if no other processes are waiting for the CPU.However, if GBL_CPU_TOTAL_UTIL is
Generated with www.html-to-pdf.net Page 15 / 17

100 percent and GBL_RUN_QUEUE is greater than the number of processors, it indicates a CPU bottleneck.
Back to top

Are there any tricks to installing a new instance of HP-UX? When a new install of HP-UX is done, set the vg00 maximum number of extents per physical volume to at least 5000.This is necessary to allow the addition of large disks to vg00 later.
Back to top

How can I get a user to have the privileges of more than one group simultaneously? If you have users that need to belong to more than one group, run the following command to give those users simultaneous access to all groups to which they belong. This is necessary at 10.20 and earlier, however the default behavior at HP-UX 11.0 is to allow such access without the /etc/logingroup file. #ln /etc/group /etc/logingroup
Back to top

How can I search all the man pages for keywords? Run the first command below to build the man pages keyword database.Access this database by using the second command below. #catman w #man k <keyword to search on>
Back to top

How can I synchronize the system clocks of all my systems? If you have systems that pass data between them, or communicate with each other in a manner where time/date information is important, set up time synchronization between all of these servers.The easiest way to do this is with the NTP facility through sam.The systems clocks must be accurate and precise.Precision of time among multiple systems is achieved by selecting one system as the NTP time master and configuring all other systems to set their system clock according to the master. Thus, the systems in an application group can maintain the same time.Time accuracy can be achieved by configuring the time master to periodically set its time according to some external known-good time source, such as the clock at NIST (National Institute of Standards and Technology) or the Naval Observatory.

time-a.nist.gov time-b.nist.gov tick.usno.navy.mil tock.usno.navy.mil

Back to top

129.6.15.28 129.6.15.29 192.5.41.40 192.5.41.41

Gaithersburg, Md. Gaithersburg, Md. Washington, DC Washington, DC

White Papers and Other Documentation


The following list identifies white papers and other documents written by Pete Gulotta for your viewing pleasure.Use your browsers Back button to return to this page after viewing any of these documents. Kernel Parameter Setting is a brief overview of the most commonly modified HP-UX kernel parameters. Standard Root Volume Group Configuration is a description of a viable, standard root disk layout. The Layered Availability Model is a paper written by Pete Gulotta and presented at the HP World 2000 seminar.It is a description of a systematic approach for building high availability. This paper, On Imagination, highlights two examples of the working of the imagination. This overview of the syslog facility describes its use and configuration options. This is a list of the HP-UX 11.11, 11.23 and 11.31 Kernel Parameters default values.Note the tabs at the bottom of the page for the three revisions. Time is a brief explanation of how HP-UX keeps and reports time.
Back to top

And another thing a.k.a. Petes Soapbox


The following is a list of papers on some of Petes favorite subjects and rants.All have been written by Pete Gulotta.These articles are meant to be both thought provoking and informative.Use your browsers Back button to return to this page after viewing any of these documents. Why Do I Need a Performance Management Strategy is a brief explanation of default and deliberate performance management strategies.
Generated with www.html-to-pdf.net Page 16 / 17

Who Should Do Performance Analysis is an exploration of the consequences of having anyone, or worse, everyone, do performance analysis and troubleshooting. Performance Reports is a commentary on two different approaches to performance reporting. Stop Working More Than 40 Hours per Week is a commentary on the excessive hours put in by IT professionals. TMI is a rant about electronic mail.
Back to top

Generated with www.html-to-pdf.net

Page 17 / 17

Vous aimerez peut-être aussi