HPUX ADMINISTRATION NOTES



HP-UX SYSTEM ADMINISTRATION NOTES

By: Russell Becker

Updated by: Court Campbell

BOOT ERROR CODES

1000’s CPU Crashing Errors

2000’s CPU Crashing Errors

3000’s CPU Crashing Errors

5000’s Internal I/O Errors hang here

7000’s Memory Errors hang here

8000’s I/O Bus Transaction Errors hang here

A Unix OS Runtime Errors Hang here

B Unix OS Runtime Errors Hang here

C Unix OS Runtime Errors Hang here

(A good resource for boot error codes is Ed Griglak 440-238-4377 at Wellington Technologies)

LOG FILES

/etc/rc.log – rc script output from current boot.

/etc/rc.log.old - rc script output from previous boot

/var/adm/syslog/syslog.log - syslogd facility messages of info or higher from current boot.

/var/adm/syslog/OLDsyslog.log – syslogd facility messages of info or higher from previous boot.

/var/adm/syslog/mail.log – syslogd mail facility messages

/logdat/maste* - ??? (Need more information)

/var/mail/root - Root’s mail. Sometimes errors are reported here.

SECURITY AND SYSTEM LOGS

/var/adm/wtmp - contains a record of all logins and logouts. To disable logging remove this file. Use the command last to view its contents. To clear the file use the following command:

# > /var/adm/wtmp

/var/adm/btmp - Contains bad login entries for each invalid logon attempt. To disable logging remove this file. Use the command lastb to view its contents. To clear the file use the following command:

# > /var/adm/btmp

/var/adm/sulog - logs use of su command

/etc/shutdownlog - logs everytime the system is shutdown

/etc/securetty – file which lists ttys from which root can log in.

/usr/adm/inetd.sec – optional security file for indetd. See man inetd.sec for more info.

Hosts.equiv, .rhosts – secureity files authorizing access by remote hosts and users on local host. See man .rhosts for more info.

|Command |Information |

|# /opt/ignite/bin/print_manifest | more |Prints a formatted ASCII system manifest. (See man |

| |print_manifest) |

|# uname -i |Displays the machines identification number. |

|# mem |Displays the amount of physical memory. |

|# cstm |Using CSTM you can query the Memory sizes and count. |

|cstm> map | |

|cstm> sel dev (memnumber) | |

|cstm> info | |

|cstm> infolog | |

|cstm> q | |

|# top |Display and update information about the top processes on the|

| |system |

|# model |On N, L, A class machines this will display the processor |

| |speed. |

|# itick_per_tick/D | adb -k /stand/vmunix /dev/kmem | tail -1 | awk |On all other systems use this string of commands to display |

|'{print $2/10000}' |the processor speed. I would suggest copying the command to a|

| |shell script and running it. |

|# /usr/sbin/ioscan –fnC tape |This will display the tape devices that the system |

| |recognizes. (DDS-2 is C1533A and DDS-3 is C1537A) |

|# /usr/sbin/ioscan –fnC lan |Type of Network Card and Driver |

|# /usr/sbin/ioscan –fnC disk |Type/ Number of Disk devices the system sees |

|# find / -exec grep “string” {} /dev/null \; |Find a string of text anywhere on a system |

|# getconf KERNEL_BITS |Displays the number of bits the kernel is, ie, 32 or 64. |

|sed ‘s/string/newstring/’filename > tempfilename |replace a line of text within a file |

|what /stand/vmunix | grep -i btlan |See if a patch is in the kernel: (ex. Looking for “btlan”) |

|/usr/contrib/bin/show_patches –a |See all patches install on system |

|ipcs –mb |See processes using shmem |

|# vxupgrade /home |See what version of JFS a filesystem is. (to upgrade to v4, |

| |use “–n 4”) |

|# swapinfo –tam |View Swap information including what is locked and vg’s |

|# kmtune | more |View current kernel settings |

|or | |

|# kmtune –q parameter | |

|# sysdef |grep bufpages |Check Buffer Cache Size. (Multiply by 4096 to get mb) |

QUERYING A SYSTEM FOR INFORMATION

|# fuser –cu /directory |View files opened by a user in a directory |

|# fuser –k |Kill processes opened by a file system |

|# fuser –cu | wc –w |Get a process number count accessing a given file system |

|# fstyp –v /dev/vgXX/rlovlX |See Default Blksize for a filesystem. Note: Defaults are HFS = 8k, VXFS = 8k |

|# /usr/bin/lsof /directory |List open (currently accessed) files. (this feature is not standard, download from the HP |

| |Porting and archive center, search for lsof.) |

|# file –c filename |See if a file with a count of 0 is really empty |

LOGIN BANNER / LOGIN PROMPT MODIFICATION

1. Banner

a. Edit the /etc/issue file with what you want to say.

b. Edit the /etc/inetd.conf file and add “-b /etc/issue” after the following line :

“telnet stream tcp nowait root /usr/lbin/telnetd telnetd”

(If you want to have a blank message, don’t follow the –b with a filename)

c. You then need to restart inetd.

2. Prompt

a. Put the following Lines in the appropriate ~/.profile, or the /etc/profile:

export HISTFILE=$HOME/.sh_history

export HISTSIZE=500

export EDITOR=vi

MONITORING SYSTEM ACTIVITY

1. top - lets you see realtime system processes, daemons, processor, and memory Information.

# /usr/sbin/top

2. GLANCEPLUS is a purchased utility that is that provides performance information.

# /opt/perf/bin/glance

You can run Glance using the GUI by exporting your display and then

# /opt/perf/bin/gpm &

3. VMSTAT Virtual memory statics

# /usr/sbin/vmstat

4. SAR (System Activity Reporter) Partially set up on HP & easy to configure, this utility can run at the command line to report on system activity and resource usage, or run through cron and set to log. Here is how you set it up to run every 20 min, log, and clean up log files older than 7 days:

Create the Directory the logs will be built in:

# mkdir /var/adm/sa

Change ownership of log directory:

# chown root:sys /var/adm/sa

Change permissions of log directory:

# chmod 775 /var/adm/sa

Set the following lines in root’s crontab:

# system activity reporting

* * /usr/lbin/sa/sa1 1200 3 * * /usr/lbin/sa/sa2 -s 00:01 -e 23:59 -i 1200 -A

Related files:

/usr/lbin/sa/sa1 - Command text file that is called to define and build the logs.

/usr/lbin/sa/sa2 - Command text file that is called to clean up files older than 7 days. (This is configurable by changing the +7.)

/usr/lbin/sa/sadc – executable that collects information from the system for SAR.

If you just want to run at the command line for current information, use the following command:

/usr/sbin/sar 1 5

(this show 5 checkpoints, 1 second apart each)

SYSTEM UTILITIES

1. STM - (Support Tools Manager) Low level diagnostic utility.

# xstm (runs it in Xwindows)

# cstm (runs it in character mode)

# mstm (runs it in menu mode. Preferred)

View Details on Specific piece of hardware:

# cstm (Enter command mode)

cstm> map

cstm> sel dev dev# (Where dev# is selected from the “map” output)

cstm> info (Sends the info to the “infolog”

cstm> infolog (Displays all info requested, Cumulative…Your data is at end of log)

cstm> q

2. EMS - (Event Monitoring Systems) EMS is the data collector and trap sending agent that STM acceses. This utility is used to cause actions (SNMP, Email, Etc) as defined when specified parameters or failure occur on the system. Everything from Filesystem sizes to Hardware down states. See Section IX for specifics. EMS is accessed in SAM under “Resource Management” ( “Event Monitoring Service”. EMS is also access by command line utility by executing:

# /etc/opt/resmon/lbin/monconfig

To List all items and their current monitored states:

#/etc/opt/resmon/lbin/set_fixed –L

To clear out the list, for the next interval: #/etc/opt/resmon/lbin/set_fixed –n \*

To disable monitoring of a specific resouce, edit the following file:

/var/stm/data/tools/monitor/disabled_instances

Example for disabling all default disk resources: /storage/events/disks/default/*

3. SET PARMS - Interactive utility used for setting and changing primary Networking information.

# set_parms ip - Use this utility to change the systems primary ipaddress

# set_parms addl - Use this utility to change the systems primary default gateway name, default gateway ip address, and subnet mask.

4. SAM - (System Administration Manager) Menu driven utility to manage most all common system configurations and functions.

To start utility #/usr/sbin/sam

5. LANDIAG - Network diagnostic utility. If you want to use it, you will first want to do a get the nmid for the interface you wish to monitor:

# lanscan

To start the utility:

# /usr/sbin/landiag (/usr/sbin/lanadmin is the same)

STARTUP

Beginning with HP­UX 10 /etc/inittab calls /sbin/rc, which in turn calls execution scripts to start subsystems. This approach follows the OSF/1 industry standard and has been adopted by Sun, SGI, and other vendors. There are four components to this method of startup and shutdown: /sbin/rc, execution scripts, configuration variable scripts, and link files.

/sbin/rc

This script invokes execution scripts based on run levels. It is also known as the startup and shutdown sequencer script.

Execution scripts

These scripts start up and shut down various subsystems and are found in the /sbin/init.d directory. /sbin/rc invokes each execution script with one of four arguments, indicating the "mode":

start Bring the subsystem up

start_msg Report what the start action will do

stop Bring the subsystem down

stop_msg Report what the stop action will do

These scripts are designed never to be modified. Instead, they are customized by sourcing in configuration files found in the /etc/rc.config.d directory. These configuration files contain variables that you can set. For example, in the configuration file /etc/rc.config.d/netconf you can specify routing tables by setting variables like these:

ROUTE_DESTINATION[0]="default"

ROUTE_GATEWAY[0]="gateway_address"

ROUTE_COUNT[0]="1"

The execution script /sbin/init.d/net sources these and other network-related variables when it runs upon system startup. More on configuration files is described below.

Upon startup a checklist similar to the one below will appear based upon the exit value of each of the execution scripts.

HP-UX Startup in progress

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

Mount file systems..............................[ OK ]

Setting hostname................................[ OK ]

Set privilege group.............................[ OK ]

Display date...................................[FAIL]*

Enable auxiliary swap space....................[ N/A ]

Start syncer daemon.............................[ OK ]

Configure LAN interfaces........................[ OK ]

Start Software Distributor agent daemon..........[ OK ]

The execution scripts have the following exit values:

0 Script exited without error. This causes the status OK to appear in the checklist.

1 Script encountered errors. This causes the status FAIL to appear in the checklist.

2 Script was skipped due to overriding control variables from /etc/rc.config.d files or for other reasons, and did not actually do anything. This causes the status N/A to appear in the checklist.

3 Script executed normally and requires an immediate system reboot for the changes to take effect. (NOTE: Reserved for key system components).

GENERAL NETWORKING

1. To view what speed your network is currently running at and what duplex use:

#lanscan

2. To see what your CRD IN# is for the card in question.

#lanadmin –s (CRD IN#)

3. To hardcode a specific rate or duplex, edit/etc/rc.config.d/hpeisadtconf

4. #lanscan - Shows all interfaces and the asscoiated NMID’s & CRD IN#’s

5. #netstat –rn - Shows the active routing table. (what the system has running now)

6. #lanscan –v - To see interfaces and their state, (similar to ifconfig –a for solaris)

7. #ifconfig lan(n) - To see the state of a specific interface.

8. #ypwhich - If it returns a hostname, NIS is running (see NIS section for more)

9. #ioscan –fnC lan - Let’s you see what kind of card(s) you have. All cards except “Built-in LAN” are 100mb capable. Cards that start with “BTLAN..” are 100 mb capable cards. This is the most common.

10. #/uusr/sbin/nettl – Network Tracing and Logging control command. See man page for more info.

ADD DEFAULT GATEWAY

#route add default 10.192.4.1

*This can be used for anything in the routing table, just substitute a different interface(instead of “default) and assign an ipaddress.

REMOVE DEFAULT GATEWAY

#route delete default 10.192.4.1

*This can be used for Anything in the routing table, just substitute a different Interface(instead of “default”) and currently assigned ip.

13. To change your primary IP address the easy way, execute

#/sbin/set_parms ip

or you can edit /etc/rc.config.d/netconf

CHANGE YOUR IP ADDRESS: (do not do this thru CDE!)

a. Bring down the current interface (lan0 for example):

#ifconfig lan0 down

b.If your subnet is changing, unplumb the interface:

#ifconfig lan0 unplumb

c. Clear old entries out of routing tables:

#netstat –f (This will drop other interfaces, if you have more than one!)

d. Change the IP address in /etc/hosts

e. Change necessary entries in /etc/rc.config.d/netconf

f. Re-Plumb the interface (ONLY if you unplumbed!):

#ifconfig lan0 plumb

g. Stop and then start network services:

#/sbin/init.d/inetd stop

#/sbin/init.d/net stop

#/sbin/init.d/net start

#/sbin/init.d/inetd start

h. Do a #netstat –rn and make sure you have only the new “default” route. If the old one is still there, route delete it out (route delete default 192.60.64.1)

DISABLE DHCP

comment out the below in the /etc/inetd.conf:

bootps dgram udp wait root /usr/lbin/bootpd

bootpd

SET UP DNS

1. Edit the line listing “hosts” in the /etc/nsswitch.conf file to look at DNS.

Example:

“hosts: files dns”

2. Edit /etc/resolv.conf if necessary to have the domain and nameserver.

Example:

domain

nameserver 10.192.4.20 #cslvtele

If you have more than one domain, try this:

search

nameserver 10.1.7.55

IMPORTANT FILES

/etc/hosts - Keeps all hostname and ip associations.

/etc/rc.config.d/netconf - File that keeps all interface information. Read at boot. Volatile!

/etc/nsswitch.conf - Active table for node lookup.Defines lookup order.(Files,DNS,NIS)

/etc/resolv.conf - Active file of DOMAIN name and NAMESERVER

/etc/hosts.equiv - System list of equivalent hosts allowing logins without passwords

/sbin/init.d/hostname - RC Script that assigns the system name at boot

DRIVER INFORMATION

NIC card drivers are typically found on Disk 1 of the Applications CD set. If you are using a release March 2000 or later, the driver name is the same as the Product # Otherwise you will have to contact HP (Website, or Phone Support) for driver information. Have the cards product number ready for reference. Some Drivers I have used.

J3514A 100baseT on the HSC Expansion Card. Drivers: J3620BA/J3514A

A3495-60101 HP.PB 10/100baseT Card. Drivers: J2759BA/A3495

A3658A EISA Card on a D class box. Still have not had success!

Bringing Up A Second Interface To Same Server:

Edit the /etc/rc.config.d/netconf file:

If the interface is acknowledged, it should be represented by a list of values in the netconf file by a trailing id number surrounded by square brackets [1]. This number should be different than the primary, usually the primary is [0] and the second is [1]. Make sure the information following the second interfaces [1] is correct.

Next in the netconf file, you want to set a default gateway for the secondary interface. If your gateway ip address is 172.16.2.1 then add the following lines, substituting the “1” in [1] for the actual id for your secondary interface if it varies.

ROUTE_GATEWAY[1]=172.16.2.1

ROUTE_COUNT[1=1

ROUTE_DESTINATION=”net 172.16.2.x”

20. EXTENDING EXISTING FILE SYSTEMS (Non ONLINE JFS)

If you want to extend /usr from 300MB to 600MB, and there is available space, do the following:

a. Determine lvol # of /usr. Type:

#bdf

Make note of the lvol that is mounted to /usr.

b. Unmount the file system.

#umount /usr

If you run into problems you may have to boot into single user mode, or you may need to kill all processes associated with /usr.

#fuser –k /usr

c. Increase lvol6 to 600MB:

#/sbin/lvextend -L 600 /dev/vg00/lvol6

d. Resize the file system to size of lvol6

#/sbin/extendfs /dev/vg00/rlvol6

e. Now mount the file system:

#/sbin/mount /dev/vg00/lvol6 /usr

f. Make sure changes took effect using bdf.

g. If you are in single user mode, type the following to get to run-level 3:

#init 3

EXTENDING A VOLUME GROUP:

1. Prepare disk(s) for Logical Volume Manager (LVM) structures:

#pvcreate /dev/rdsk/???

2. Use vgextend to add the disk to the volume group:

#vgextend /dev/vg00 /dev/dsk/c?d?s?

EXTENDING ONLINE JFS FILESYSTEMS

If you want to extend /usr from 300MB to 600MB, and there is available space, do the following:

1. Determine lvol # of /usr. Type:

#bdf

Make note of the lvol that is mounted to /usr.

2. Increase lvol6 to 600MB:

/sbin/lvextend -L 600 /dev/vg00/lvol6 (-L is new size in MB)

3. Extend the file system using OnlineJFS:

#/sbin/fsadm –F vxfs –b 600m /usr (m is for MB)

In this example you want to extend /usr from 300MB to 600MB but there is no available space in the volume group that the /usr filesystem is on. In this case you would extend the volume group, then the logical volume, and finally the filesystem.

1. If adding disk online (disk array), create device files:

#insf –e (verify #ioscan –fnC disk)

2. Prepare disks for Logical Volume Manager (LVM) structures:

#pvcreate /dev/rdsk/???

3. Make space available for volgroup:

#vgextend /dev/vg00 /dev/dsk/c?d?s?

4. Increase lvol6 to 600MB:

/sbin/lvextend -L 600 /dev/vg00/lvol6 (-L is new size in MB)

5. Extend the file system using OnlineJFS:

#/sbin/fsadm –F vxfs –b 600m /usr (m is for MB)

REDUCING AN ONLINE JFS FILESYSTEMS

If you want to shrink /usr from 600MB to 300MB and the filesystem is already less than 300MB:

1. Determine lvol # of /usr using bdf.

2. Defragment and cluster the data:

#fsadm –F vxfs –d –D –e –E /mountpoint

3. Shrink the file system:

#fsadm –F vxfs –b 300000 /mountpoint (-b Stated in KB)

4. Shrink the Logical Volume:

#/sbin/lvreduce -L 300 /dev/vg00/lvol6 (-L is new size in MB)

5. Check the filesystem size using bdf.

CREATING NEW VOLUME GROUPS AND LVOL’s BY COMMAND LINE

1. Physically add the hard drive and bring the server back up.

Use the command ioscan –fnC disk and find the associated device file name for the drive.

2. Create the physical volume

#pvcreate /dev/rdsk/cXtXdX

3. Make the new LVM directory. For this example we will create vg02.

#mkdir /dev/vg02

#mknod /dev/vg02/ group c 64 0x020000

*When doing mknod, the 2 numbers are the major and minor. The major number is always 64 in this case. The minor number relates to the decimal to hexadecimal conversion of the volume group number. This is more of a guideline than a rule.

4. Create the volume group

#vgcreate –s 8 /dev/vg02 /dev/dsk/cXtXdX (if the disk used to contain LVM, use a “-f”)

5. Now that you have created the volume group you can create the first lvol:

#lvcreate –L 400 –n lvol1 /dev/vg02 (This will create a 400mb filesystem called lvol1)

#/usr/sbin/lvcreate -L 1536 -i 2 -I 8192 -n lvolname /dev/vg02 (Striped w/ 8mb PE)

*You cannot mirror/stripe internal drives, or you will get the following error:

Striped mirrors are not supported. To enable mirroring options (-m, -M, -c), do not specify the striping options (-i, -I) when creating logical volumes.

IMPORTING VOLUME GROUPS (ACTIVATE-DEACTIVATE)

We will import volume group vg02 from disk c2t2d0 in this example:

#vgimport –v /dev/vg02 /dev/dsk/c2t2d0

Next activate the volume group:

#vgchange –a y /dev/vg02 (the “a” is for “activate” and the “y” is for a state of active)

Now the logical volumes can be mounted to your desired mountpoints. If you are not sure how many logical volumes are there, just use the “vgdisplay” command:

#vgdisplay –v /dev/vg02 | more

To deactivate the volume group (So you can export it, or mount elsewhere), use this command:

#vgchange –a n /dev/vg02 (the “n” changes the state to inactive.)

To backup a volume group do the following:

/usr/sbin/vgcfgbackup –f /savefilename.vgcfg /dev/vgXX

To save a specific vg map before vgexporting, use the following command:

/usr/sbin/vgexport -pvs -m /map.vgXX /dev/vgXX

To completely remove the volume group from the system, use the following command:

#vgexport –f /tmpfile –v /dev/vg02

*The “-f” option sends the LV information to a file, that can later be used as an input file when importing it later. This will save you from having to recreate the lvols when you import it again. The “–v” is for verbose output.

ADDING ADDITIONAL DISK DEVICES (using SAM)

1. Down the box and physically add the drive(s)

2. When you come up, verify it in SAM at Disks and Filesystems -> Disk Devices

3. If you want to add the space using SAM, go to Disks and Filesystems -> Disk Devices

4. Select the drive and go to Actions -> Disk Array Maintenance -> Bind LUN

5. Enter the amount you want to use, select “yes” . When it finishes scanning, proceed w/6

6. Select the newly added hardware path. Then go to Actions -> Add

7. You will be prompted to choose LVM or non-LVM. Select LVM.

8. A new drop down menu shuold appear, Create should be highlighted, hit Return.

9. Select a valid VG#, then select OK.

10. You will go back to the drop down menu, you can add LVOL’s now, or tab to “OK”

REMOVING DISK DEVICES USING SAM

1. Select the drive and go to Actions -> Disk Array Maintenance -> Unbind

2. Answer “yes” if you are sure. Remember, all data will be lost, no going back if “yes”

3. Down the box and remove the hardware if you wish.

Note: When trying to extend existing filesystems, if it says there is no unsed space, but you know there is unsed space on a given device, do the following:

1. Go to Disks and Filesystems -> Disk Devices and select the device in question.

2. Go to Actions -> Disk Array Maintenance -> Bind LUN

Here you should see available un-bind space.

ADDING SWAP SPACE (Command Line)

1. Create new swap:

lvcreate –L 1024 –C y –r n –n /dev/vg04/lvol1/swap2 (Makes 1024mb contigous bad block off, called swap2 on vg04/lvol1)

swap on /dev/vg04/swap2 (Creates swap with default priority of 1)

2. Verify it is active:

#swapinfo –tam

3. Add to /etc/fstab file

CREATE NEW SWAP WITH STRIPING

lvcreate –L 1024 –i 4 -I 8192–r n –n /dev/vg04/lvol1/swap2 (Makes 1024mb w/ 4 stripes of 8192 in size. bad block off, called swap2 on vg04/lvol1)

MANAGING “LARGEFILES”

“Largefiles” are 2048mb or larger. By default, an HPUX install does not support this. You need to check the status of the filesystem in question, then issue some commands and edit the fstab file to enable.

1. Checking to see if a filesystem is “largefiles” enabled:

#/usr/sbin/fstyp –v /dev/vgXX/lvolX

If this returns a value of “16” for f_flag, then “largefiles” are enabled, and you can exceed 2048mb. If the f_flag is set to “0” then you have a 2047mb Maximum per file.

2. To enable “Largefiles” remove the “no” from “nolargefiles” in the /etc/vfstab for the filesystem. Then issue the command to change:

#/usr/sbin/fsadm –F vxfs –o largefiles /dev/vgXX/lvolX (You can specify the mount point if you are using Online JFS)

Finally unmount and mount the filesystem

CREATING A BOOT DISK – ROOT VOLUME

1. First create a bootable logical volume disk.

#pvcreate –B /dev/rdsk/c0t1d0

2. Add the PV to vg00.

#vgextend /dev/vg00 /dev/dsk/c0t1d0

3. Place Boot Utilities on disk:

#mkboot /dev/rdsk/c0t1d0

4. Add the AUTO File to the LIF area:

#mkboot –a “hpux –lq (;0)/stand/vmunix” /dev/rdsk/devicefile

MIRRORING

Definitions

RAID Levels

RAID 0 - Speed

Level 0 is disk striping only, which interleaves data across multiple disks for better performance. It does not provide safeguards against failure. RAID 0 is widely used in gaming machines for higher speed.

RAID 1 - Fault Tolerance

Uses disk mirroring, which provides 100% duplication of data. Offers highest reliability, but doubles storage cost. RAID 1 is widely used in business applications.

RAID 2 - Speed

Bits (rather than bytes or groups of bytes) are interleaved across multiple disks. The Connection Machine used this technique, but this is a rare method.

RAID 3 - Speed and Fault Tolerance

Data are striped across three or more drives. Used to achieve the highest data transfer, because all drives operate in parallel. Parity bits are stored on separate, dedicated drives.

RAID 4 - Speed and Fault Tolerance

Similar to Level 3, but manages disks independently rather than in unison. Not often used.

RAID 5 - Speed and Fault Tolerance

Data are striped across three or more drives for performance, and parity bits are used for fault tolerance. The parity bits from two drives are stored on a third drive and are interspersed with user data. RAID 5 is widely used on servers to provide speed and fault tolerance.

RAID 6 - Speed and Fault Tolerance

Highest reliability, but not widely used. Similar to RAID 5, but performs two different parity computations or the same computation on overlapping subsets of the data.

RAID 10 - Speed and Fault Tolerance

A combination of RAID 1 and RAID 0 combined. Raid 0 is used for performance, and RAID 1 is used for fault tolerance.2.

MIRRORING THE ROOT VOLUME – NEED TO EDIT THIS

1. First create a bootable logical volume disk. (See CREATING A BOOT DISK – ROOT VOLUME)

2. Mirror the root logical volume to new disk:

#lvextend –m 1 /dev/vg00/lvol1 /dev/dsk/c0t1d0

3. Mirror the Primary Swap area:

Lvextend –m 1 /dev/vg00/prswaplv /dev/dsk/c0t1d0

4. Verify the LVM was updated:

#lvlnboot –v

BREAKING A MIRROR FOR BACKUP PURPOSES

Mirrors are created and broken from an lvol level. Remember this. If you want to break the mirror for /dev/vg00/lvol1 which may be mounted at / do the following:

#lvsplit /dev/vg00/lvol1

This will by default create a new, unmounted lvol called lvol1b. You could also do the following:

#lvsplit –s backup /dev/vg00/lvol1

This does the same, but names it lvol1backup

Mount the new lvol to what ever you want, and back it up.

REMIRRORING AFTER BACKUP

After you are finished, you must remirror and sync with existing and new data created on The active mirror. Use this command (using the lvol1b from example above):

#lvmerge /dev/vg00/lvolb /dev/vg00/lvol1

REMOVING A MIRROR (SO YOU CAN PULL A DRIVE)

1. You must first remove the association between each lvol associated with the drive. Use the following command to view the current mirrored lvol’s:

#lvlnboot –v

2. Break the mirror and remove the pv/lvol association for each lvol.

This example removes PV /dev/dsk/c1t0d0 from /dev/vg00/lvol2, and lvol3. Assuming that there was only 1 other disk in the mirror set:

#lvreduce –m 0 /dev/vg00/lvol2 /dev/dsk/c1t0d0

#lvreduce –m 0 /dev/vg00/lvol3 /dev/dsk/c1t0d0

MAIL CONFIGURATION (SENDMAIL)

1.Setting up mail aliases

#vi /etc/mail/aliases

Enter aliases after the line “Local aliases”. Here is an example:

#Local aliases

usergroup :user@

othergroup :localuseraccount

2. Important files

/usr/spool/mqueue - Directory that mail spools up in.

/etc/mail/sendmail.cf - Config file.

/etc/mail/sendmail.cw - List of all hostnames that are local.

3. Get Sendmail version

#what /usr/sbin/sendmail

NTP (Network Time Protocol)

1. Important files:

/etc/rc.config.d/netdaemons - Daemon file.Enables/Disables XNTPD from running.

/etc/ntp.conf - NTP configuration file. Should contain the NTP Server IP Address, driftfile path(/etc/ntpdrift), and controlkey.

/etc/ntp.keys - file contains “key”code, or specific number that is required to authenticate the NTP server .

/etc/ntp.drift - File controling the amount time allowed to drift from the control server before it is corrected.

2. To set up NTP on an HP BOX:

a. Change /etc/TIMEZONE if necessary to TZ=UTC0

Set current Timezone, so you do not have to reboot.

#TZ=UTC0;export TZ

#ntpdate –d 192.50.54.129 (compare current time to central time server)

b. Sync time with NTP sever:

#ntpdate 192.50.54.129

c. Edit the /etc/rc.config.d/netdaemons file, set “export XNTPD=1”. NOT “0”

d. Edit the /etc/ntp.conf file, enter new line “peer 192.50.54.129” for time server.

e. Start the xntpd daemon.

#/usr/sbin/xntpd start

PRINTING

1. Install “HP Printer Installer” Software (replaced Jetadmin)

A. Download software from HP site :

B. Follow install.unx instructions

C. Once installed use the /opt/hpnpl/bin/hppi utility to manage/create printers.

2. Management

A. #lpstat - Shows the status of all requests still in the output queue.

B. #lpsched – Starts the scheduler, then runs in the background. (It is OK to issue this command, even if scheduler is already running.)

C. #lpshut – Stops the scheduler. All printers that are printer will stop, requests will restart when lpsched is issued.

D. #enable – Allows printers to accept requests from LP

E. #disable – Disallows printer from accepting requests from LP

F. #strings /var/spool/outputq | more - Show listing of all printer queues

G. #lpadmin [options] - This command has many options and is used to setup and change printer configurations. After printers are initially set up with “hppi” it is often easier to use SAM

3. Import Files/Directories

/var/adm/lp/log – LP activity and error log.

/etc/lp/interface – Files created that “are” the printers.

/opt/hpnpl/admin – Directory that contains commands for further, more complex pinter configurations and setup.

/var/spool/lp/interface/model.orig – Directory that contains the Model files (Drivers) for each printer that has been set up.

/opt/hpnp/tmp/printername.log or /opt/hpnpl/tmp/printername.log – Printer log.

4. Printer Problems

A. There was no "Stopping" scheduler message, it looks like the scheduler was messed up inadvertently. To fix this, you need to delete the following 2 files, /var/spool/lp/FIFO and /var/spool/lp/SCHEDLOCK. lpsched should then start up fine.

B. If you are unsure if a remote printer’s Jet Direct is an internal card, or external box, just

telnet to the printer and if the Firmware Revision starts with J,H,B,or E, it is External.If it starts with A,D,G,K,or L then it is Internal.

C. When deleting a printer, if you try to remove the printer and it says it cannot because there are jobs, just remove the files in line 2 above, and then issue the following commands:

#more /dev/null > /var/spool/lp/outputq

#more /dev/null > /var/spool /lp/qstatus

#more /dev/null > /var/spool/lp/pstatus

#rm –r /var/spool/lp/request/printername/*

D. When you create a transferqueue from a system that is currently running HPPI Only, any printers created with Jetadmin may show unusual problems, such as stopping large print jobs short. To fix this problem, recreate the printer within the HPPI utility. If this still fails, you will need to make a transferqueue backup, remove the HPPI software, add Jetadmin, add HPPI again, and then transferqueue the printers back in. After you do this, make a backup again, then another transferqueue. This will clean up the .old files in /etc/lp/interface/model.orig. Then copy the model script (/etc/lp/interface/model.orig/printer) from the server it came from.

E. JetAdmin and HPPI support, call The Jetdirect Printer Group at 208-323-2551 Menu 3,1,1

F. Banner Page Controls:

If you want to turn off the banner page, you can either set it in HPPI or manually edit the file /etc/lp/interface/model.orig/PRINTERNAME. Look for the corresponding line value (It is case sensitive! Use all caps, it is the value that counts):

BANNER=”yes”

BANNER=””

If the printer has a jetdirect, Telnet to the jetdirect, and enter the command where the number is the quantity of banner pages you wish the jetdirect to produce:

banner:0

CHANGE A PRINTER TO USE A4

cd /var/spool/lp/interface/model.orig

vi # this will be the name of the printer

edit the line that says paper=”def” to say paper=”ISOA4”

XTERM

1. To export a display. On the server you are sending from, enter the following command:

#export DISPLAY=172.3.10.29:0 (IP address an expample) the :0 is the first display you export. :1, :2, etc are the next ones you export.

2. Next, run the command you want exported. EX: /etc/appname & (run in background).

On the server that is receiving, you may get an error saying connection refused, or something..

To fix this, type one of the following commands:

#xhost + (allows all connections.) Add and ip address(s) after the +n and it will include only the ipaddress(s)specified.

#xhost – (excludes all connections) Add an ip address(s) after the – and it will exclude only the specified ip address(s).

SOFTWARE & PATCHES

1. Commands

A. #swlist –l fileset | more

or

#/usr/contrib/bin/show_patches -a

Lists all patches on server

B. #swlist –l product | grep –i phco_10947

Checks the system to see if patch is in the kernel.

C. #swlist –l fileset –a state | grep phco_10947

Checks to see if patch is installed or configured.

Note: Patches that return “installed” are not active, usually because dependencies have not been met. You can change it’s state to “configured” by running the swconfig command:

#swconfig –x mount_all_filesystems=false PHCO_10947

Also you should check to see if the patch has any errors by using the “-a patch_state” option. Patches that are error free show a state of “Configured” and a patch_state of “Applied”:

#swlist –l fileset –a state –a patch_state | grep phco_10947

#check_patches - This lets you see if the system has any patch dependencies not yet satisfied.

This software (command) is not available unless it is loaded. It is patch number PHCO_24347. To get patches off the Internet, log on to . Use user ID “CA395965” and password is “f0rg0tit”. Select your patch and download directly to an HP server/workstation.

2. Installing a patch.

a. Shar the patch:

#sh PHCO_10947 (this decompresses the patch)

b. Open up the file PHCO_10947.text and search for swinstall.

c. Install patch with this command:

#swinstall -s /tmp/package.depot \*

3. Creating a depot (Bundle of patches, or software) so you can install multiple patches at once.

a. Make a new directory to be used as the depot.

For example: mkdir /tmp/depot

b. Copy each shared patch or SW into the new depot directory by using “swcopy” :

#swcopy -s /tmp/PHCO_10947.depot \* @ /tmp/depot

Register the depot

#swreg –l depot /tmp/depot

To install the newly bundled Depot (or any depot, you can use the swinstall utility by typing:

#swinstall, or by the command line by using this command:

#swinstall -s /tmp/depot \*

4. To see what PATCH LEVEL you are at, look at the largest number directory in the /tmp/PHTXT directory. An Additional way is to look for the 2 files below when you do a

#swlist –l product | more

ExtSWDocs B10.20.41 Bundle Readme Files

PatchText B10.20.41 Patch .text Files

5. Patches have the ability to not be “committed”, thus they can be swremoved if you need. The only issue is that it consumes a lot of space in /var/adm/sw/save. If you need to free up space use the following command to commit patches you think you will not need to ever back out.

#swmodify –x patch_commit=true PHCO_12345.\*

Some patches are installed, but not configured. To configure all patches, run this command:

#swconfig \*

If sam fails to start, check the SAM patches with the following command:

#swlist –l fileset –a state SystemAdmin*

This will show all SAM patches, make sure they are all Configured. If not, run thE following command:

#swconfig SysAdmin*

FTP

1. For FTP to work, several conditions must be met:

A. The username attempting login must be in the /etc/passwd file.

B. The username must NOT appear in the /etc/ftpusers file.

C. There must me a line in the /etc/inetd.conf file specifying to start ftpd within inetd:

Example:

ftp stream tcp nowait root /usr/lbin/ftpd ftpd –l

D. There must be a port assigned to ftp in the /etc/services file:

Example:

ftp-data 20/tcp #File Transfer Protocol (Data)

ftp 21/tcp #File Transfer Protocol (Control)

2. Exception files:

/etc/shells - Include file that contains the path of the shells allowed during ftp sessions.

/etc/ftpusers - Exclude file used to lock users out.

SYSTEM FAILURES

1. Crashes & Panics Usually described as such in /etc/shutdownlog

A. Go to the crash Directory (Defined in /etc/rc.config.d/savecrash: default /var/adm/crash):

# cd /var/adm/crash

B. Find the latest crash, and cd into it.

# ls –al crash* (the newest dated directory is what you want. Ex. crash.1)

# cd crash.1

C. Unzip the kernel and image files. Make sure you have room!!! Type the following commands:

# q4pxdb vmunix (preps the file to be diagnosed)

# q4 (starts the diagnostic utility)

# trace event 0 (Brief description of events that caused the panic)

If it shows an “HPMC” crash event it is a hardware related issue. Usually CPU mem Cache problems. If it does not show HPMC, check for LPMC.

D. While in crash directory, set Q4 Envrionment:

#. /usr/contrib/Q4/bin/set_env

E. Cut and paste the following to the command line. (entire block)

q4 –p . he - Displays the help menu.

GSP> he li -when “he” is used before any command, it display info about it.

GSP> li – Lists the commands you can use.

GSP> lc – gives and sets the LAN configuration

GSP> ls – lists LAN configuration

GSP> so – Changes security , login and access options.

Your default login and password are blank. Just hit the enter key on new systems, and then use “so” to configure the login information.

2. Clearing the attention light on front of L and N class servers

a. GSP> sl (This take you to view the errors. This also clears the attention light. If you log in to the gsp via IP address in telnet, it will print the following message after successful login:

[Read only – use ^Ecf for console write access]

Get around this by typing “CTRL + e” followed by the letter “c” and finally “f”.

Virtual Array (VA) Disk Array

1. When first connecting the array, install special files:

#insf –e

2. Search for connected arrays:

#/opt/sanmgr/commandview/client/sbin/armdiscover

3. Identify the serial number:

#/opt/sanmgr/commandview/client/sbin/armdiscover –v

4. Start the management utility by exporting the display so a gui will run, then execute:

#/opt/sanmgr/commandview/client/sbin/cmdviewVA serialnumber &

FC60 & Autoraid 5447A Disk k

1. Commands for Autoraid

A. arraydsp –a serial number(give detail information of Autoraid unit and and state)

B. arraydsp –i (gets serial number)

C. logprint –a serial number detailed log print of autoraid activities.Redirect to file)

D. amcfg –L B:5 –d 2:1,4:1,6:1 –r 5 FC60_1 (Adds controller B to disks 2:1,4:1,6:1 running raid 5, to array alias FC60_1.

E. amdsp –r arrayid Checks if any luns are currently rebuilding.

F. amdsp –R arrayed Scans for new disk arrays.

G. amcfg –M 12 –c B FC60_1 (Switches primary controller for Lun 12 to B on FC60_1)

2. Daemons

/sbin/init.d/hparray [start|stop] - Start this daemon first for Autoraid & FC60

/sbin/init.d/hparamgr [start|stop] - Start this after the above for the FC60.

If it will not start, and you get a message in the syslog “software logging suspended”, then you need to clear out the logs. Do so by the following steps, and then start it:

# cd /var/opt/hparray/log

# rm L00000*

# rm AMMCATLG

# rm LOGCATLG

# touch AMMCATLG

# chmod 444 AMMCATLG

# chown root:root AMMCATLG

CRON AND CRONTAB

1. Important Files

/var/adm/cron - Main Cron Directory

/var/adm/cron.allow - List of allowed users

/var/adm/cron.deny - List of denied users

/var/adm/cron/log - Accounting log file for cron

/var/spool/cron/crontabs - Directory that keep the actual cron files, the files are each named after the user who owns the crontab.

2. Commands

A./usr/bin/crontab –l - List the entries in the current users crontab

B./usr/bin/crontab –e - Use the default editor to edit the crontab for the current user

This is the way the crontab should be altered, it checks for errors in format, and I if errors are found, it will not save the current crontab. It also forces a reread by by the cron daemon. It is normally not recommended to manually edit the /var/spool/cron/crontabs

C. /usr/bin/crontab –r - Remove the crontab for the current user

3. File Entries

A. The entries in the crontab must be listed in the following order:

Minute hour monthday month weekday command

PFS Mounting Oracle CD’s

1. As the root user, start the following daemons:

#/usr/sbin/pfs_mountd &

#/usr/sbin/pfsd &

#/usr/sbin/pfs_mount /dev/dsk/c3t2d0 /SD_CDROM (assuming SD_CDROM exists

2. Change user to Oracle with the following command. Do not vary!!!!

#su oracle

3. Copy the contents of the CD:

#cp –R /SD_CDROM/* /destination_directory

4. When finished, change user back to root, and unmount CD

#exit

#/usr/sbin/pfs_umount /SD_CDROM

5. Load the next CD and start over

Remotely Mounting a CDROM

1. Local Host (Server CD is physically on)

A. Start the pfs daemon

# nohup /usr/sbin/pfs_mountd & (start pfs_mount daemon)

# nohup /usr/sbin/pfsd & (starts pfs daemon)

B. Add the line “/cdrom” to the file /etc/pfs_fstab

C. Add the line “/cdrom -ro” to the file /etc/pfs_exports

D. # pfs_exportfs –a (Tells daemon to export everything in the pfs_* files)

#pfs_mount /dev/dsk/devicefile /cdrom

E. Remote Host (Server you want to remotely mount it on, or access it from)

#nohup /usr/sbin/pfs_mountd &

#nohup /usr/sbin/pfsd &

F. This command may not be necessary if you do a –a on 1E:

pfs_mount –o xlat=unix systemwherecdis:/cdrom /cdrom

Adding Memory to a System

1. N-Class Servers

A. Memory Sizes May Be Mixed, but sizes must be paired (i.e. 0a+0b same size)

B. When Mixing Sims In the Same Extender, Use largest Pairs first:

0a=1024 1a=512 2a=256

0b=1024 1a=512 2a=256 (and so on...)

Tape Changer Control Commands (DLT)

Mount tape in slot 4 to drive:

#mc -p /dev/dltpt -s S4 -d D1

Dismount tape in drive and move back to slot 4:

# mc –p /dev/dlpt –s D1 –d S4

To control a HP DLT Auto Changer device, use the “uma” command. Do a man on UMA for options. A command I have written down that I have used is:

#/opt/omni/lbin/uma –ioctl /dev/picker

where /dev/picker is the control device for the robotic arm of the drive. Once in the utility there will be a prompt that is the same as your control file name. The following commands will help you:

stat - Show status of Drives and Tapes

move S1 D1 - Move tape from Slot 1 to Drive 1.

move D2 S1 – Move tape drive 2 to Slot 1

Copy Printers from Machine to Machine

1. On the host machine run the following command:

# /usr/sam/lbin/lpmgr -S -xsavedir=/var/tmp/printers.copy

2. Tar the directory using the following command:

# tar –cvf /var/tmp/printers.copy.tar /var/tmp/printers.copy

3. Use scp to copy the file to the target machine:

# scp /var/tmp/printers.copy.tar @:/var/tmp/

4. Extract the tar on the remote machine

# tar –xvf /var/tmp/printers.copy.tar

5. Install the printers file to the new system and:

# /usr/sam/lbin/lpmgr -R -xsavedir=/var/tmp/printers.copy

UNSORTED Notes

1. If you want to edit the /etc/passwd file, use /usr/sbin/vipw (it checks for errors)

2. If you want to create a user by command line, use the following:

# useradd –m –c “Comments here” username

After you have created the user you will have to set the password manually. If you want to force the user to change the password you just set, after the first login, put a comma and 2 dots after the data in the password field of the /etc/passwd file. (ie: username:encryptedpw,..:)

3. To check and assign device files to a system (not intrusive if they already exist) use this command:

#insf –e

4. To view volume groups and associated device files:

#strings /etc/lvmtab

5. If you are having difficulty searching really large man pages, use this command to strip out reverse display:

#man command | col –b > /tmp/commandtemp

6. Need to remove a file with an invalid or invisible filename?

# ll -i | more (Note: Verify the file's inode.)

# find . -xdev -inum # -exec ll {} \; (Note: Verify ll of the inode.)

# find . -xdev -inum # -exec rm -i {} \; (Note: Confirm the deletion of each file.)

ACRONYMS

IFS = Internal Field Seperator

SAM = System Administration Manager

ISL = Initial System Loader

IPL = Initial Program Loader

LVM = Logical Volume Manager

LP = Line Printer

TZ = Time Zone

I/O = Input / Output

TCP/IP = as

NFS = Network File System

ACL = Access Control List

HP VUE = Hewlett Packard Visual User Environment

CDE = Common Desktop Environment

CRT = Cathode Ray Tube

RAID = Redundant Arrays of Independent Disks

STM = Support Toll Manager

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download