SUN Solaris Tips

4/3/12 SUN Solaris Tips

1/24www.akadia.com/services/solaris_tips.html

Akadia Information Technology

SUN Free SoftwarePCNFS installierenInstallation Solaris mit Openwindows (Grafikkarte)How to Backup a SystemPackages (Software die installiert wurde)Monitor Mode (OK Prompt)Kernel AnalyseDefaults einstellenWichtige KonfigurationsfilesAdmin KommandosLAN konfigurierenIP-Routing konfigurierenDNS konfigurierenAnonymous FTP aufsetzenNFS-Client KonfigurationNFS-Server KonfigurationAutomounterModem konfigurierenSCSI-Harddisk an SUN Hardware anschliessenList Solaris Hardware ConfigurationShow Swap Space currently installedShow Operating System Patch LevelHow to install a Sun Solaris Jumbo Patch ?Tracing System CallsTroubleshooting Solaris Device FilesShort Tips to maintain Sun SolarisIP-Aliasing for SUN SolarisSolaris automounter installs filesystems by default in /netSolaris keyboard utilityMonitoring PerformanceEnable file system journaling on Solaris 7 and 8Solaris Syslog Daemon DebuggingDoes each Oracle Process use more than 100M memory ?Sizing up Solaris Memory with the RMCmem PackageUsing Sun Solaris Manuals directly from CD-ROMWhy is the Sun Solaris System Corefile helpful ?DLT-TAPE UNIT INSTALLATION on Solaris 7/8/9Reconfigure Devices on SolarisOpenBoot DiagnosticsWhy doesn't my .forward file workSimple Shell Script to backup your Files

SUN Free Software

Unter http://www.sunfreeware.com findet man "ready to use" Software für SUN Solaris, wie beispielsweise TOP,AMANDA, GCC, GDB etc. Download via FTP von: «ftp://nce.sun.ch/pub/freeware/sparc/7»

PCNFS installieren

CD-ROM Solaris Intranet Extension (siehe auch Solaris Server Intranet Extension Installation)

$ su$ cd /cdrom/cdrom0/nfsc/sparc$ pkgadd -d pwd

Installation Solaris mit Openwindows (Grafikkarte)

Hostname

$ uname -u

Network Interface

$ ifconfig -a

/etc/hosts, /etc/netmasks definieren

Static IP-routes definieren



Static IP-routes definieren

/etc/rc2.d/S79staticroutes

CD-ROM rausnehmen

$ eject cdrom

Disklayout kontrollieren

$ prtvtoc /dev/rdsk/....

Automounter konfigurieren

/etc/auto_master, /etc/auto_home

Device File für DAT

/dev/rmt/0l (tar cvf /dev/rmt/0l)

/etc/system definieren konfigurieren (Prestoserve, Oracle, Transtec)

Logfile der Installation: /var/sadm/system/logs/install_log

Installation über serielles Terminal an Nullmodem Kabel

OK boot cdrom - w (Terminal an ttya)

How to Backup a System

$ init 0OK boot -s$ fsck -m /dev/dsk/c0t0d0s0 (und übrige Filesysteme)$ tar cvf /dev/rmt/0l

Packages (Software die installiert wurde)

Anzeige der installierten Packages

$ pkginfo

Check ob Package SUNWpcnfd correct installiert ist

$ pkgchk -v SUNWpcnfd

Package installieren (Path ist meistens /cdrom/cdrom0/....)

$ pkgadd -d <Path to CDROM> SUNWpcnfd

Das Package SUNWpcnfd entfernen

$ pkgrm SUNWpcnfd

Monitor Mode (OK Prompt)

In single user mode booten

OK boot -s

Kernel zwingen /devices neu aufzubauen nach dem Anschluss von neuer Hardware

OK boot -r

Detaillierter Bootvorgang

OK boot -v

Vom CD-ROM aus booten: Notboot !

OK boot cdrom

Angeschlossene SCSI-Geräte testen

OK probe-scsi



List System Devices, e.g. SUNW,hme = Sun Fast Ethernet PCI Adapter

OK show-devs

List Network Devices

OK show-nets

Monitoring Network Activity

OK apply watch-net <full path name of the hme interface, see show-devs>

Monitor Variablen ändern, anzeigen

OK eepromOK eeprom ttya-mode=38400,8,n,1,h

Kernel Analyse

Welche Kernel-Module sind geladen ?

$ modinfo

Kernel Konfiguration

/etc/system

Logfile von syslog

/var/adm/messages

Konfiguration des syslog Daemon

/etc/syslog.conf

Defaults einstellen

Directory mit Default files

/etc/default

Remote root logins erlauben

/etc/default/login

Timezone setzen

/etc/default/init

Wichtige Konfigurationsfiles

Master-File beim Booten

/etc/inittab

Run-Level Start/Stop Files

/etc/rc?.d

Scripts für Run-Levels

/etc/init.d

Admin Kommandos

$ shutdown -g0 -i0$ reboot (entspricht init 6)

LAN konfigurieren

Konfiguration der LAN-Interfaces

$ ifconfig -a

Netmask setzen: siehe /etc/netmasks



Netmask setzen: siehe /etc/netmasks

Jedes LAN-Interface hat /etc/hostname.le0 mit Hostnamen

LAN-Setup: /etc/rcS.d/S30rootusr.sh (Interfaces konfigurieren)/etc/rc2.d/S72inetsvc (LAN konfigurieren)

Phys Addressen nachschauen

$ arp -a

Net to Media Table

Device IP Address Mask Flags Phys Addr------ -------------------- --------------- ----- ---------------

le0 rabbit 255.255.255.255 00:60:08:57:17:86le0 quorum 255.255.255.255 SP 08:00:20:89:27:03le0 arkum 255.255.255.255 00:a0:24:4b:60:1c

IP-Routing konfigurieren

Alle hosts im Netz 193.72.239.0 werden über den Router 193.72.194.201 erreicht.

$ route add net 193.72.239.0 193.72.194.201 1

Der host 146.228.14.10 wird über den Router 193.72.194.100 erreicht. Siehe File /etc/rc2.d/S79staticroutes.

$ route add host 146.228.14.10 193.72.194.100 1

Routing Tabelle kontrollieren

$ netstat -nr

DNS konfigurieren

Angabe des DNS Nameservers

/etc/resolv.conf

Reihenfolge definieren

/etc/nsswitch.conf

Anonymous FTP aufsetzen

Siehe Solris2 Administration Seite 103 und ff

NFS-Client Konfiguration

/etc/vfstab (Soll) --> /etc/mnttab (Ist)mount -F nfs -o bg,ro,soft gondwana:/usr/software /software

NFS-Server wird in /etc/init.d/nfs.client start gestartet.

Anzeige welche Directories gondwana zum mounten freigegeben hat

dfshares gondwana

RESOURCE SERVER ACCESS TRANSPORTgondwana:/export/home/zahn gondwana - -gondwana:/export/home/steiner gondwana - -

NFS-Server Konfiguration

/etc/dfs/dfstab (Soll) --> /etc/dfs/sharetab

Directory read-only freigeben

$ share -o ro /usr/software

Alle Directories in /etc/dfs/dfstab freigeben

$ shareall

Alle Directories in /etc/dfs/dfstab zurücknehmen



Alle Directories in /etc/dfs/dfstab zurücknehmen

$ unshareall

NFS-Server wird gestartet in

/etc/init.d/nfs.server

Anzeige der freigegbenen lokalen Direcories

$ share

Anzeige welche Clients nutzen welche Directories eines NFS-Servers

$ dfmounts -F nfs gondwana

RESOURCE SERVER PATHNAME CLIENTSgondwana /export/home/zahn paragon.glue.ch,rabbit.glue.ch

Automounter

- /etc/auto_master (Master Map konfigurieren)- /etc/auto_home (Home Direcories verwalten)- autofs ist ein spezielles Filesystem- automount -v (Nach einer Aenderung an einer Map ausführen)

Modem konfigurieren

Siehe spezielles Dokument

SCSI-Harddisk an SUN Hardware anschliessen

Beispiel: SCSI-Disk Seagate ST150176L, 50MB an SUN Ultra Enterprise 1

Eintrag in /etc/format.dat vornehmen (Angaben von Lieferanten)

disk_type = "Seagate ST150176L" \: ctlr = "SCSI" \: ncyl = 12022 : acyl = 2 : pcyl = 12024 : nhead = 22 : nsect = 369 \: rpm = 7200 : bpt = 188928

Eintrag /etc/system für Solaris-2 Kernel, System booten

** SCSI-Disc Konfiguration*set scsi_options=0x20

Disk anschliessen, SCSI-Adresse kontrollieren, Terminierung

Unbedingt kontrolieren, dass eine SCSI-Adresse nicht mehrfach belegt ist. Dazu kann meistens hinten amGerät ein Tippschalter eingestellt werden. Man beacht, dass in der Regel das letzte Gerät terminiert werdenmuss.

Disk formatieren (nur wenn notwendig !)

In der Regel muss eine Disk nicht neu formatiert werden, ist dies jedoch notwendig so steht unter Solarisdas Utility format zur Verfügung.

format

AVAILABLE DISK SELECTIONS:

0. c0t0d0 <SUN2.1G cyl 2733 alt 2 hd 19 sec 80> /sbus@1f,0/espdma@e,8400000/esp@e,8800000/sd@0,01. c0t1d0 <SUN2.1G cyl 2733 alt 2 hd 19 sec 80> /sbus@1f,0/espdma@e,8400000/esp@e,8800000/sd@1,02. c0t2d0 <IBM-DDRS-39130-S71D cyl 8186 alt 2 hd 10 sec 218> /sbus@1f,0/espdma@e,8400000/esp@e,8800000/sd@2,03. c0t4d0 <SEAGATE-ST118273N-5764 cyl 7499 alt 2 hd 20 sec 237> /sbus@1f,0/espdma@e,8400000/esp@e,8800000/sd@4,04. c0t5d0 <SEAGATE-ST150176LW-0002 cyl 12022 alt 2 hd 22 sec 369> /sbus@1f,0/espdma@e,8400000/esp@e,8800000/sd@5,0

Specify disk (enter its number): 4



Specify disk (enter its number): 4

format> type

AVAILABLE DRIVE TYPES:0. Auto configure1. DDRS-391302. Seagate ST1182733. Seagate ST150176L4. Quantum ProDrive 80S5. Quantum ProDrive 105S6. CDC Wren IV 94171-3447. SUN01048. SUN02079. SUN032710. SUN034011. SUN042412. SUN053513. SUN066914. SUN1.0G15. SUN1.0516. SUN1.3G17. SUN2.1G18. SUN2.9G19. IBM-DDRS-39130-S71D20. SEAGATE-ST118273N-576421. SEAGATE-ST150176LW-000222. otherSpecify disk type (enter its number)[21]: 21

format> format (confirm with "yes")

Disk partitionieren

Dadurch wird die Disk in logische Teile unterteilt. Jeder teil enthält ein eigenes Filesystem.

format> part

Nun die Partitionierungsdaten eingeben, zB

partition> print

Current partition table (original):

Total disk cylinders available: 2733 + 2 (reserved cylinders)

Part Tag Flag Cylinders Size Blocks0 root wm 0 - 204 152.15MB (205/0/0) 3116001 swap wu 205 - 377 128.40MB (173/0/0) 2629602 backup wm 0 - 2732 1.98GB (2733/0/0) 41541603 home wm 378 - 1017 475.00MB (640/0/0) 9728004 unassigned wm 0 0 (0/0/0) 05 unassigned wm 1018 - 1928 676.13MB (911/0/0) 13847206 usr wm 1929 - 2732 596.72MB (804/0/0) 12220807 unassigned wm 0 0 (0/0/0) 0

Label erzeugen (aktuelle Partitionierung speichern)

partition> labelpartition> ypartition> quitformat> quit

Filesystem erstellen

newfs -v -m 0 /dev/rdsk/c0t5d0s0

Damit wird ein Filesystem mit 0 % Min-Free auf der Partition 0 der Disk an der SCSI-Adresse 5 erstellt.

Filesystem mounten

Dazu den folgenden Eintrag in /etc/vfstab vornehmen

#device device mount FS fsck mount mount#to mount to fsck point type pass at boot options



/dev/dsk/c0t5d0s0 /dev/rdsk/c0t5d0s0 /u02 ufs 6 yes -

List Solaris Hardware Configuration

$ /usr/sbin/prtconf

Show Swap Space currently installed

Multiply the Blocks column by 512

$ swap -l

swapfile dev swaplo blocks free/dev/dsk/c0t0d0s1 32,1 16 262944 262944

262944 * 512 = 134 MB

Show Operating System Patch Level

$ showrev -p

Patch: 105181-16

Note, that Patchlevel 105181-15 is minimal needed for Oracle 8.1.6

How to install a Sun Solaris Jumbo Patch ?- Download the Patch from: http://sunsolve.sun.com- Read the README File included in the Patch- Usually the only thing you have to do is:

$ cd <patch cluster directory>$ ./install_custer$ cat /var/sadm/install_data/<luster name>_log$ showrev -p

Reboot the system

Tracing System Calls

You can trace system calls with truss on Solaris an strace on Linux

$ truss svrmgrl

Troubleshooting Solaris Device Files

If you suspect troubles with your Solaris device files, e.g. system doesn't boot after a filesystem check, you mayrepair the solaris system using the following commands.

Halt the system immediately with the keys STOP-A, you will now see the boot prompt: OK

STOP-A

Reset the machine with

OK reset

Boot the machine with

OK boot -r

The command boot -r will rebuild all devices files according to your attached hardware. If you cannot boot themachine, you can try the following commands: drvconfig, disks, tapes

drvconfig - configure the /devices directory

The default operation of drvconfig is to create the /devices directory tree that describes, in the filesystemnamespace, the hardware layout of a particular machine. Hardware devices present on the machine and poweredon as well as pseudo-drivers are represented under /devices. Normally this command is run automatically after anew driver has been installed (with add_drv(1M)) and the system has been rebooted.

disks - creates /dev entries for hard disks attached to the system

Disks creates symbolic links in the /dev/dsk and /dev/rdsk directories pointing to the actual disk device special



Disks creates symbolic links in the /dev/dsk and /dev/rdsk directories pointing to the actual disk device specialfiles under the /devices directory tree.

tapes - creates /dev entries for tape drives attached to the system

Tapes creates symbolic links in the /dev/rmt directory to the actual tape device special files under the /devicesdirectory tree. Tapes searches the kernel device tree to see what tape devices are attached to the system.

Short Tips to maintain Sun Solaris

Here are some short tips for common tasks on SUN Solaris 2.6, 7 and 8

Important SUN Solaris Commands

$ who -r # Show Run Level$ /usr/sbin/prtconf # Print the complete system configuration$ /sbin/mountall -l # Mount all local filesystems.$ /sbin/init S # Changing to single user mode

Show currently mounted filesystems

# /etc/mnttab: Contains information about devices that# are currently mounted. If there are mounted filesystems# with quotas enabled, display them

if /usr/bin/cut -f 4 /etc/mnttab | \ /usr/bin/egrep 'quota|,quota' >/dev/null 2>&1; then echo 'There are mounted filesystems with quotas enabled'fi

How to enable system activity data gathering

# You will also need to uncomment the sa entries in# the system crontab /var/spool/cron/crontabs/sys. # Refer to the sar(1) and sadc(1m) man pages# for more information.

$ /usr/bin/su sys -c "/usr/lib/sa/sadc /var/adm/sa/sadate +%d"

How a new, unused Solaris system is setup ?

# sysidtool is a suite of five programs that configure a new# system, or one that has been unconfigured with sys-# unconfig(1M). The sysidtool programs run automatically at# system installation, or during the first boot after a# machine has been successfully unconfigured.## These programs have no effect except at such times, and# should never be run manually.

# System Files are

cat /etc/nodenamecat /etc/hostname.*cat /etc/default/initcat /etc/defaultdomaincat /etc/inet/hostscat /etc/inet/netmasks

How to configure Asynchronous PPP ?

Configure /etc/asppp.cf for the aspppd daemon

$ /usr/sbin/aspppd -d 1

How to get and set TCP/IP driver configuration parameters ?

# Getting Parameters Supported By The TCP Driver# To see which parameters are supported by the TCP driver,# use the following command:

$ ndd /dev/tcp \?

# The following command sets the value of the parameter# ip_forwarding in the IP driver to zero. This disables IP



# ip_forwarding in the IP driver to zero. This disables IP# packet forwarding.

Disable IP Forwarding

$ /usr/sbin/ndd -set /dev/ip ip_forwarding 0

Enable IP Forwarding (Machine acting as a Router)

$ /usr/sbin/ndd -set /dev/ip ip_forwarding 1

How to set Default Route on Solaris ?

# Configure default routers using the local "/etc/defaultrouter"# configuration file. The file can contain the hostnames or IP# addresses of one or more default routers.## The default routes listed in the "/etc/defaultrouter" file will# replace those added by the kernel during diskless booting. An# empty "/etc/defaultrouter" file will cause the default route# added by the kernel to be deleted.## Note that the default router file is ignored if we received routes# from a DHCP server. Our policy is to always trust DHCP over local# administration.

# Set Default Route

$ route -n add default <Default Route from /etc/defaultrouter>

# Show Default Route

$ /usr/sbin/route -fndefault 128.128.128.11 done

How to set NIS domainname if locally configured ?

if [ -f /etc/defaultdomain ]; then /usr/bin/domainname cat /etc/defaultdomain echo "NIS domainname is /usr/bin/domainname"fi

RPC (Remote Procedure Call) Configuration

# rpcbind - universal addresses to RPC program number mapper# rpcinfo - report RPC information

Solaris Keyserv Daemon

# keyserv is a daemon that is used for storing the private# encryption keys of each user logged into the system. These# encryption keys are used for accessing secure network ser-# vices such as secure NFS and NIS+.

$ /usr/sbin/keyserv

How to start the Solaris DNS server "in.named"

# If this machine is configured to be an Internet # Domain Name System (DNS) server, run the name daemon.# Start named prior to: route add net host,# to avoid dns gethostbyname timout delay for# nameserver during boot.

if [ -f /usr/sbin/in.named -a -f /etc/named.conf ]; then echo 'starting internet domain name server.' /usr/sbin/in.named &fi

Where to find syslogd messages ?

Configuration File: /etc/syslog.confMessage File: /var/adm/messages

IP-Aliasing for SUN Solaris



IP-Aliasing for SUN Solaris

# How to setup IP-Alias on SUN Solaris

1. Setup File /etc/hostname.hme0:1 for the second IP-Address

cat /etc/hostname.hme0:1

ldap

2. Insert IP-Address in /etc/hosts

# # Internet host table # 128.128.128.11 ux-portal1 # IP-address on hme0:0 128.128.128.20 ldap # IP-alias on hme0:1

3. Start alias IP-Address on Interface in /etc/rc2.d

S99ipalias -> ../init.d/ipalias

#!/bin/sh # Akadia AG, Arvenweg 4, CH-3604 Thun # ---------------------------------------------------------------------- # File: ipalias # # Autor: Martin Zahn / 10.05.2000 # # Purpose: Setup second IP address on hme0:1 # ----------------------------------------------------------------------

if [ -f /etc/hostname.hme0:1 ] then case "$1" in

'start') # Start second IP address on hme0:1

echo "Start multi-homed server for UX-ALIAS1 on hme0:1" ifconfig hme0:1 128.128.128.20 up ;;

'stop') # Stop second IP address on hme0:1

echo "Stop multi-homed server for UX-ALIAS1 on hme0:1" ifconfig hme0:1 128.128.128.20 down ;; esac fi

4. Check IP-Address on second Interface

ifconfig -a

Solaris automounter installs filesystems by default in /net

The Solaris automount utility installs autofs mount points and associates an automount map with each mount

point. The autofs file system monitors attempts to access directories within it and notifies the automountd

daemon. The daemon uses the map to locate a file system, which it then mounts at the point of reference withinthe autofs file system. You can assign a map to an autofs mount using an entry in the /etc/auto_master map

or a direct map in /etc/auto_direct. If the file system is not accessed within an appropriate interval (five

minutes by default), the automountd daemon unmounts the file system.

Default Mapping under /net

The mount point /net is by default the location, where automountd mounts NFS filesystems, which are exported

on other machines. Lets suppose, that you have the filesystem /home exported on the NFS server saphir, then

the (Solaris) NFS client with an active automounter will automatically mount this NFS filesystem under/net/saphir/.

Mapping using /etc/auto_direct

You probably doesn't want this default behavior. If you insert the following entry in /etc/auto_direct ....

/opt/local -rw remote_machine:/local



.... then, the directory /local on the remote machine "remote_machine" will be mounted on the local machine

under /opt/local.

Solaris keyboard utility

The Solaris utility kbd manipulates the state of the keyboard or display the type of keyboard or change the

default keyboard abort sequence effect. Suppose, that you do not want that everybody can halt the thesystem you must change the default value. We also noticed, that the Solaris machines attached to a switch box,using a character terminal on a serial line, may halt when you switch from one machine to the other.

SYNOPSIS

kbd [ -r ] [ -t ] [ -c on|off ] [ -a enable|disable ] [ -d keyboard device ]kbd -i [ -d keyboard device ]

DESCRIPTION

kbd manipulates the state of the keyboard, or displays the keyboard type or allows the default keyboard abort

sequence effect to be changed. The default keyboard device being set is /dev/kbd.

The -i option reads and processes default values for the keyclick and keyboard abort settings from the keyboarddefault file, /etc/default/kbd. Only keyboards that support a clicker respond to the -c option. If you want to

turn clicking on by default, add or change the current value of the KEYCLICK variable to the value on in thekeyboard default file, /etc/default/kbd, as shown here.

KEYCLICK=on

Then, run the command 'kbd -i' to change the current setting. Valid settings for this variable are the values on

and off. Other values are ignored. If the variable is not specified in the default file, the setting is unchanged.

The keyboard abort sequence (L1-A or STOP-A) on the keyboard and BREAK on the serial console input deviceon most systems) effect may only be changed by the superuser, using the-a option. On most systems, the default effect of the keyboard abort sequence is to suspend the operatingsystem and enter the debugger or the monitor.

If you want to permanently change the software default effect of the keyboard abort sequence, you can add orchange the current value of the KEYBOARD_ABORT variable to thevalue disable in the keyboard default file, /etc/default/kbd, as shown here.

KEYBOARD_ABORT=disable

Then, run the command 'kbd -i' to change the current setting. Valid settings for this value are the values

enable and disable. Other values are ignored. If the variable is not specified in the default file, the setting isunchanged.

OPTIONS

-i Set keyboard defaults from the keyboard default file. This option is mutually exclusive withall other options except for the -d keyboard device option. This option instructs thekeyboard command to read and process keyclick and keyboard abort default values from the/etc/default/kbd file. This option can only be used by the superuser.

-r Reset the keyboard as if power-up.-t Return the type of the keyboard being used.-c On/Off state Turn the clicking of the keyboard on or off.-a Enable/Disable state; Enable or disable the keyboard abort sequence effect.

Monitoring Performance

This chapter describes procedures for monitoring system performance by using the vmstat, iostat, df, and

sar commands. This is a list of the step-by-step instructions in this chapter.

How to Display Virtual Memory Statistics (vmstat)

The following example shows the vmstat display of statistics gathered at five-second intervals.

$ vmstat 5

procs memory page disk faults cpur b w swap free re mf pi po fr de sr f0 s3 -- -- in sy cs us sy id0 0 8 28312 668 0 9 2 0 1 0 0 0 1 0 0 10 61 82 1 2 970 0 3 31940 248 0 10 20 0 26 0 27 0 4 0 0 53 189 191 6 6 880 0 3 32080 288 3 19 49 6 26 0 15 0 9 0 0 75 415 277 6 15 79



0 0 3 32080 256 0 26 20 6 21 0 12 1 6 0 0 163 110 138 1 3 960 1 3 32060 256 3 45 52 28 61 0 27 5 12 0 0 195 191 223 7 11 820 0 3 32056 260 0 1 0 0 0 0 0 0 0 0 0 4 52 84 0 1 99

Category FieldName

Description

procs Reports the following states:

r The number of kernel threads in the dispatch queue

b Blocked kernel threads waiting for resources

w Swapped out LWPs waiting for processing resources to finish

memory Reports on usage of real and virtual memory:

swap Available swap space

free Size of the free list

page Reports on page faults and paging activity, in units per second:

re Pages reclaimed

mf Minor and major faults

pi Kbytes paged in

po Kbytes paged out

fr Kbytes freed

de Anticipated memory needed by recently swapped-in processes

sr Pages scanned by page daemon (not currently in use). If sr does not equal zero,the page daemon has been running.

disk Reports the number of disk operations per second, showing data on up to fourdisks

faults Reports the trap/interrupt rates (per second):

in Interrupts per second

sy System calls per second

cs CPU context switch rate

cpu Reports on the use of CPU time:

us User time

sy System time

id Idle time

How to Display System Event Information

Run vmstat -s to show the total of various system events that have taken place since the system was last

booted.

0 swap ins 0 swap outs 0 pages swapped in 0 pages swapped out409376480 total address trans. faults taken 3075036 page ins 2601555 page outs 3812452 pages paged in 6525552 pages paged out 11007609 total reclaims 10927650 reclaims from free list 0 micro (hat) faults409376480 minor (as) faults 2957386 major faults102738273 copy-on-write faults 61711047 zero fill page faults1002562077 pages examined by the clock daemon 7881 revolutions of the clock hand 16716370 pages freed by the clock daemon

4999048 forks 1138206 vforks 5747009 execs



5747009 execs741660225 cpu context switches736047593 device interrupts528054538 traps2496638575 system calls430283487 total name lookups (cache hits 95%) 81727 toolong 10484677 user cpu 9528364 system cpu443762786 idle cpu 16281790 wait cpu

How to Display Swapping Statistics

Run vmstat -S to show swapping statistics.

procs memory page disk faults cpu r b w swap free si so pi po fr de sr m1 m3 m4 m5 in sy cs us sy id 0 0 0 8512 888 0 0 12 21 55 0 417 1 0 0 0 206 1040 308 2 2 96

si = Average number of LWPs swapped in per secondso = Number of whole processes swapped out

How to Display Disk Utilization Information (iostat)

You can display disk activity information by using the iostat command with a time interval. The following

example shows disk statistics gathered every five seconds.

iostat 5

tty md1 md3 md4 md5 cpu tin tout kps tps serv kps tps serv kps tps serv kps tps serv us sy wt id 0 2 10 1 28 2 0 22 0 0 0 1 0 10 2 2 3 92 0 47 58 7 39 16 2 34 0 0 0 0 0 0 0 2 19 78 0 16 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 98 0 16 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 99 0 16 2 0 22 0 0 0 0 0 0 0 0 0 2 3 1 95 0 24 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 98

For Each ... Field Name Description

Terminal

tin Number of characters in the terminal input queue

tout Number of characters in the terminal output queue

Disk

bps Blocks per second

tps Transactions per second

serv Average service time, in milliseconds

CPU

us In user mode

sy In system mode

wt Waiting for I/O

id Idle

How to Display Extended Disk Statistics

Run iostat -xtc to get extended disk statistics. This command displays a line of output for each disk.

extended device statistics tty cpudevice r/s w/s kr/s kw/s wait actv svc_t %w %b tin tout us sy wt idmd1 0.4 0.9 3.6 6.9 0.0 0.0 27.7 1 1 0 2 2 2 3 92md3 0.1 0.2 1.0 1.3 0.0 0.0 21.7 0 0md4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0md5 0.0 0.0 0.7 0.0 0.0 0.0 9.9 0 0md8 0.8 0.3 6.7 14.2 0.0 0.0 13.1 0 1md10 0.2 0.9 1.8 6.8 0.0 0.0 15.5 0 1md11 0.2 0.9 1.8 6.8 0.0 0.0 14.8 0 1md30 0.0 0.2 0.5 1.3 0.0 0.0 11.4 0 0md31 0.0 0.2 0.5 1.3 0.0 0.0 10.2 0 0md40 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0



md40 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0md41 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0 0md50 0.0 0.0 0.4 0.0 0.0 0.0 9.4 0 0md51 0.0 0.0 0.4 0.0 0.0 0.0 7.3 0 0md80 0.4 0.3 3.3 14.2 0.0 0.0 10.3 0 0md81 0.4 0.3 3.3 14.2 0.0 0.0 11.7 0 1sd0 0.6 2.1 6.0 22.8 0.0 0.0 16.3 0 3sd1 0.6 2.1 6.0 22.8 0.0 0.0 15.2 0 2

Field Name Description

r/s Reads per second

w/s Writes per second

Kr/s Kbytes read per second

Kw/s Kbytes written per second

wait Average number of transactions waiting for service (queue length)

actv Average number of transactions actively being serviced

svc_t Average service time, in milliseconds

%w Percentage of time the queue is not empty

%b Percentage of time the disk is busy

How to Check CPU Utilization (sar)

Display CPU utilization with the sar -u command. (The sar command without any options is equivalent to sar -u.)At any given moment, the processor is either busy or idle. When busy, the processor is in either user or systemmode. When idle, the processor is either waiting for I/O completion or "sitting still" with no work to do.

Measure CPU utilization during 5 secs one time.

sar -u 5 1

Measure CPU utilization during 60 secs 1440 times and write result in file sar.log.

sar -u -o sar.log 60 1440

To later review disk and tape activity from that period:

sar -d -f sar.log

Field Name Description

%sys Lists the percentage of time that the processor is in system mode

%user Lists the percentage of time that the processor is in user mode

%wio Lists the percentage of time the processor is idle and waiting for I/O completion

%idle Lists the percentage of time the processor is idle and is not waiting for I/O

A high %wio generally means a disk slowdown has occurred.

Enable file system journaling on Solaris 7 and 8

Solaris 7 and 8 include a native implementation of file system journaling. This feature, known as "intent logging"or just "logging" enables FASTER file system operation and FASTER system boot.

It's trivial to implement and safe to use. The new logging feature is an option to the Unix File System (UFS),which is the standard file system for all disk partitions on SUN servers, except for partitions holding swap space.By default, the journaling option is disabled. Logging is enabled on a per file system basis, and it can even beenabled on / (root file system) and other operating system partitions.

Background

Solaris UFS logging works by allocating space from the file system's free blocks. Within that space, all metadatachanges to the file system are written. Metadata includes directory and I-node information but not file datablocks, essentially everything but the actual data within the file. So, for example, a "file create" modifies thedirectory structure and allocates a new I-node, and those activities are written to the logging space. Once themetadata changes are made to the logging area, the system is free to perform other operations to the filesystem. In the background, the information in the log is flushed to the file system and updates the appropriatedirectory and I-node structures, completing the file system operations.

The logging data is written sequentially within the log space. It's therefore much faster for the operating systemto complete metadata changes via logging and background flushing than by directly modifying the metadata (viarandom I/O) spread across the disk. The size of the logging space is based on the size of the file system, and



random I/O) spread across the disk. The size of the logging space is based on the size of the file system, andequals 1 MB per 1 GB of file system space, up to 64 MB. The space is used as a circular log: if the log space isabout to fill up, new metadata change requests are paused while the log is emptied. As changes are moved fromthe log to the file system, that log space is made available, and new metadata changes can be written to thelogging space.

Usually with UFS, if the system crashes during any file system operation, the entire system must have itsconsistency checked via the fsck command. That command can take several minutes per file system because itchecks all metadata and file data to ensure the structures are correct, free, and used, and that the I-node blockcounts are correct. It also confirms that the free space available is current, repairs inconsistencies, andoccasionally requires manual intervention to fix large problems. Files and even directories can be lost, dependingon the operations occurring at the time of the crash.

Because metadata changes are made first to the log space rather than to the file system, the consistency checkfor a logged file system after a crash is a simple and fast operation. The system evaluates the logging data anddetermines which changes had completed against the underlying file system, which had yet to start, and whichwere in progress. Those completed or not yet started are removed from the log, and those partly completed areeither undone or completed. If there's sufficient data in the log to complete the operation, it's completed.Otherwise, the changes made are removed from the underlying file system.

People familiar with database operation will recognize the similarity between database transaction processing andthe activities here. The end result is that the underlying file system is consistent, and no thorough consistencychecking is needed. That operation completes in a few seconds per file system.

Using logging

Starting with Solaris 7, there's a new logging option to the mount command and in the /etc/vfstab systemconfiguration file. Logging only appears in a couple other places within Solaris. The mount command shows whichpartitions are mounted and lists logging in the options fields for each partition on which logging is enabled. Finally,at system boot time, the fsck phase reports per partition whether each is stable, logging, or being checked.There are no other status commands available to determine the state of logging.

A = Device to mountB = Device to fsckC = Mount pointD = Filesystem TypeE = Fsck pass (unimportatnt with logging)F = Mount at bootG = Mount options

# ------------------------------------------------------------------# A B C D E F G# ------------------------------------------------------------------fd - /dev/fd fd - no -/proc - /proc proc - no -/dev/dsk/c0t0d0s3 - - swap - no -/dev/dsk/c0t0d0s0 /dev/rdsk/c0t0d0s0 / ufs 1 no logging/dev/dsk/c0t0d0s6 /dev/rdsk/c0t0d0s6 /usr ufs 2 no logging/dev/dsk/c0t0d0s1 /dev/rdsk/c0t0d0s1 /var ufs 3 no logging/dev/dsk/c0t0d0s7 /dev/rdsk/c0t0d0s7 /home ufs 4 yes logging/dev/dsk/c0t0d0s5 /dev/rdsk/c0t0d0s5 /opt ufs 5 yes logging/dev/dsk/c0t8d0s0 /dev/rdsk/c0t8d0s0 /u01 ufs 6 yes logging/dev/dsk/c0t9d0s0 /dev/rdsk/c0t9d0s0 /u02 ufs 7 yes logging/dev/dsk/c0t10d0s0 /dev/rdsk/c0t10d0s0 /u03 ufs 8 yes logging/dev/dsk/c0t11d0s0 /dev/rdsk/c0t11d0s0 /u04 ufs 9 yes logging/dev/dsk/c0t12d0s0 /dev/rdsk/c0t12d0s0 /u05 ufs 10 yes logging/dev/dsk/c1t13d0s0 /dev/rdsk/c1t13d0s0 /app ufs 11 yes logging/dev/dsk/c1t14d0s0 /dev/rdsk/c1t14d0s0 /users ufs 12 yes loggingswap - /tmp tmpfs - yes -

Logging increases performance, decreases fsck time, removes the risk of a file system corruption, can be used onall UFS partitions (including root), and is free.

Solaris Syslog Daemon Debugging

The log system messages daemon syslogd reads and forwards system messages to the appropriate log filesand/or users, depending upon the priority of a message and the system facility from which it originates. Theconfiguration file /etc/syslog.conf controls where messages are forwarded. The syslogd daemon ignores anyfaulty entry in /etc/syslog.conf, specially spaces instead of tabs are not recognized by syslogd. Therefore alwayscheck the entries in /etc/syslog.conf in the debugging mode of syslogd.

How to check /etc/syslog.conf

# /etc/init.d/syslog stop



# /etc/init.d/syslog stop# /usr/sbin/syslogd -d

getnets() found 1 addresses, they are: 0.0.0.0.2.2amiloghost() testing 193.247.121.196.2.2cfline(*.err;kern.notice;auth.notice /dev/sysmsg)cfline(*.err;kern.debug;daemon.notice /var/adm/messages)cfline(mail.info;mail.debug /var/log/maillog)

syslogd: line 14: unknown priority name "debug /var/log/maillog"

cfline(*.alert;kern.err;daemon.err operator)cfline(*.alert root)cfline(*.emerg *)cfline(user.err /dev/sysmsg)cfline(user.err /var/adm/messages)cfline(user.alert root, operator)cfline(user.emerg *)

syslogd: version 1.70Started: Sat Jan 6 10:11:47 2001Input message count: system 0, network 0# Outputs: 10

5 3 3 3 5 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 X CONSOLE: /dev/sysmsg7 3 3 5 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 X FILE: /var/adm/messagesX X 6 X X X X X X X X X X X X X X X X X X X X X X UNUSED:3 1 1 3 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 X USERS: operator1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 X USERS: root0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 X WALL:X 3 X X X X X X X X X X X X X X X X X X X X X X X CONSOLE: /dev/sysmsgX 3 X X X X X X X X X X X X X X X X X X X X X X X FILE: /var/adm/messagesX 1 X X X X X X X X X X X X X X X X X X X X X X X USERS: root, operatorX 0 X X X X X X X X X X X X X X X X X X X X X X X WALL:

Per File StatisticsFile Tot Dups Nofwd Errs---- --- ---- ----- ----/dev/sysmsg 0 0 0 0/var/adm/messages 0 0 0 0 0 0 0 0operator 0 0 0 0root 0 0 0 0WALL 0 0 0 0/dev/sysmsg 0 0 0 0/var/adm/messages 0 0 0 0root,operator 0 0 0 0WALL 0 0 0 0

syslogd: restartedoff & running....sys_poll blocking, init_cnt=0

# D# /etc/init.d/syslog start

Line 14 in /etc/syslog.conf are filled up with spaces instead of tabs. Replace the spaces with tabs and syslogdwill accept the new entry in Line 14.

Does each Oracle Process use more than 100M memory ?

If you check the oracle process with the OS comand "pmap" or "top", you can see that each oracle process usemore than 100M memory. Is this a problem on the Oracle installation or something else? It seems that pmapcounts the SGA size as the private memory segment of each oracle process, but we believe the SGA size shouldbe shared.

Output from "top" on our Solaris System with Orcale 8.1.7.0

PID USERNAME THR PRI NICE SIZE RES STATE TIME CPU COMMAND

-----------------------------------------------------------------361 oracle 258 59 0 124M 88M sleep 0:01 0.00% oracle373 oracle 11 59 0 122M 88M sleep 41:50 0.00% oracle363 oracle 11 59 0 119M 88M sleep 0:01 0.00% oracle



363 oracle 11 59 0 119M 88M sleep 0:01 0.00% oracle365 oracle 11 58 0 119M 88M sleep 0:17 0.02% oracle359 oracle 1 59 0 119M 89M sleep 0:00 0.00% oracle377 oracle 1 59 0 119M 88M sleep 0:00 0.00% oracle375 oracle 1 58 0 119M 88M sleep 0:00 0.00% oracle367 oracle 1 58 0 118M 89M sleep 0:00 0.00% oracle371 oracle 1 58 0 118M 89M sleep 0:00 0.00% oracle369 oracle 1 58 0 118M 88M sleep 0:00 0.00% oracle

Memory Allocation for Oracle Processes

On many UNIX platforms and specially on Sun platforms, the text of the Oracle binary and shared libraries areactually shared between background processes if these instances share the same ORACLE_HOME. So you need tosubtract the shared text of the oracle binary and the shared libraries in the result of the OS commands.

Even pmap and pmen utilities make mistakes between these memory divisions, and sometimes SGA and textexecutable are often added incorrectly.

Determine the memory used by each Oracle background process on a Solaris

This can be used by anyone who has privleges for the pmap, which can be found in /usr/proc/bin/. First, weneed to find the process id (PID) of the Oracle background process you wish to determine the memory size for.This is done by issueing the following command:

# ps -u oracle -f

UID PID PPID C STIME TTY TIME CMDoracle 359 1 0 12:26:17 ? 0:00 ora_pmon_DIA3oracle 361 1 0 12:26:17 ? 0:01 ora_dbw0_DIA3oracle 363 1 0 12:26:17 ? 0:01 ora_lgwr_DIA3oracle 365 1 0 12:26:17 ? 0:18 ora_ckpt_DIA3oracle 367 1 0 12:26:17 ? 0:01 ora_smon_DIA3oracle 369 1 0 12:26:17 ? 0:00 ora_reco_DIA3oracle 371 1 0 12:26:17 ? 0:00 ora_snp0_DIA3oracle 373 1 0 12:26:17 ? 41:50 ora_s000_DIA3oracle 375 1 0 12:26:17 ? 0:00 ora_d000_DIA3oracle 377 1 0 12:26:18 ? 0:00 ora_d001_DIA3

Second, you then enter the following commands for the DB Writer process (ora_dbw0_DIA3) with process id =361 as an example.

# /usr/proc/bin/pmap 361 | grep "shmid"80000000 82992K read/write/exec/shared [ shmid=0x2 ]

# /usr/proc/bin/pmap 361 | grep "total"total 124232K

Then you take the total size: 124232K and subtract the SGA size which the line marked with "shmid=" above, inthis case it is 82992K. So, 124232K minus 82992K is 41240K. So, the DBWR background process is approximately41.2 MB. Repeat this steps for all the background processes.

Sizing up Solaris Memory with the RMCmem Package

How much memory is needed on SUN Solaris? Explaining memory in Solaris by reviewing the different types ofmemory and introducing a set of tools, the RMCmem package.

Install RMCmem Package

Download the RMCmem tools available from ftp://playground.sun.com/pub/memtool. The package includes a kernelmodule that provides extra instrumentation.

# cd /tmp# zcat RMCmem3.8.2.tar.gz | tar xvf -# pkgadd -d .

The package is installed in /opt/RMCmem (see README in this directory)

Virtual / Physical Memory Usage

Solaris is a virtual memory system. The total amount of memory that you can use is increased by adding swapspace to the system. If you ever see "out of memory" messages, adding swap space is the usual fix. Performanceof the system is very dependent on how much physical memory (RAM) you have. If you don't have enough RAMto run your workload, performance degrades rapidly.

Physical memory usage can be classified into four groups:



Kernel memory mapped into kernel address space

Process memory is mapped into a process address space

Filesystem cache memory that is not mapped into any address space

Free memory that is not mapped into any address space

RMCmem includes a simple command to summarize this:

# /opt/RMCmem/bin/prtmem

Total memory: 989 MegabytesKernel Memory: 60 MegabytesApplication: 110 MegabytesExecutable & libs: 42 MegabytesFile Cache: 757 MegabytesFree, file cache: 11 MegabytesFree, free: 6 Megabytes

Total physical memory

The total physical memory can be seen using prtconf. Memory is allocated in units called pages, and you can usethe 'pagesize' command to see the size in bytes per page:

# /usr/sbin/prtconf | grep MemoryMemory size: 1024 Megabytes

# /usr/bin/pagesize8192

Kernel memory

Kernel memory is allocated to hold the initial kernel code at boot time, then grows dynamically as new devicedrivers and kernel modules are used. Kernel tables also grow dynamically, unlike some older versions of Unix. Asyou add hardware and processes to a system, the kernel will grow. In particular, to keep track of all the memoryin a system, the kernel allocates a page table structure.

If you have several gigabytes of RAM this table gets quite large. The dynamic kernel memory allocator grabsmemory in large "slabs," then allocates smaller blocks more efficiently. This means that the kernel tends to grab abit more memory than it's really using. If there is a severe memory shortage, the kernel unloads unused kernel

modules and devices and frees unused slabs. The simplest summary of kernel memory usage comes from sar. Toshow the kernel memory allocation (KMA) activities use (see man sar for more details).

# sar -k 1

SunOS diamond 5.7 Generic_106541-12 sun4u 04/28/01

sml_mem alloc fail lg_mem alloc fail ovsz_alloc fail6873088 6044236 0 44818432 43761720 0 11231232 0

Application process memory

Application processes consist of an address space divided into segments, where each segment maps either to afile, anonymous memory (the swap space), System V shared memory, or a memory mapped device. The mappedfiles include the code and initialized data for the command and all its shared libraries.

What we really want to know, is the amount of RAM used by each segment. This is shown by the pmemcommand in the RMCmem package.

# /opt/RMCmem/bin/pmem 361

361: ora_dbw0_DIA3 Kbytes Resident Shared Private Permissions Mapped File 82992 82992 82992 - read/write/exec [shmid=0x2] 16 16 8 8 read/exec libc_psr.so.1 16 16 8 8 read/exec libmp.so.2 8 8 8 - read/write/exec libmp.so.2........ .. .. . ............... ........... 112 80 72 8 read/exec libelf.so.1 8 8 8 - read/write/exec libelf.so.1 16 16 8 8 read/exec libkvm.so.1 8 8 8 - read/write/exec libkvm.so.1-------- ------ ------ ------ ------ 124232 93040 92728 312



Now we can see that the process address space size is 124232 kilobytes; 93040 kilobytes of that are currentlyresident in main memory, wherein 92728 kilobytes are shared with other processes while 312 kilobytes areprivate. When this command started only the 312 kilobytes of private memory were taken from the free list.

If we now go through all the processes on the system, add up how much private memory they use, and also addin the shared memory for each mapped file, we'll know how much application memory is in use. This summary isshown by prtmem as we saw in the beginning, and the detail is listed by the memps command in RMCmem.

# /opt/RMCmem/bin/memps

PID Size Resident Shared Private Process... ....... ...... ...... .... .............359 118904k 93608k 92800k 808k ora_pmon_DIA3367 118184k 93152k 92704k 448k ora_smon_DIA3369 117928k 93120k 92704k 416k ora_reco_DIA3371 118040k 93136k 92720k 416k ora_snp0_DIA3365 119040k 93120k 92712k 408k ora_ckpt_DIA3377 118344k 93080k 92720k 360k ora_d001_DIA3363 119088k 93056k 92720k 336k ora_lgwr_DIA3375 118344k 93048k 92720k 328k ora_d000_DIA3361 124232k 93040k 92728k 312k ora_dbw0_DIA3373 121608k 93032k 92728k 304k ora_s000_DIA3

Filesystem cache memory

This is the part of memory that is most confusing, as it is invisible. You can only tell it's there if you access thesame file twice and it is quicker the second time.

The RMCmem package adds kernel instrumentation that counts up all the pages for each cached file. The memps-m command lists the files that are cached in order of the amount of memory they're consuming.

One problem is that within the kernel, the file is only known by its inode number and filesystem mount point. Thedirectory pathname for the file may not be known.

The RMCmem package tries to solve this problem by catching file names as files are opened (by interposing on thevnode open code) and making an inode-to-name lookup cache in the kernel. This cache size is limited (to 8192entries by default), and the file may have been opened before the kernel module was loaded, so it can't alwaysfind the name.

# memps -m

Size InUse E/F Filename

21064k 21064k F /usr (inode 540488) 8184k 824k F /usr (inode 260922) 7752k 7752k F /usr (inode 540429) 7480k 7480k F /usr (inode 540428) 7480k 7480k F /usr (inode 540427) 6896k 6896k F /usr (inode 540450) .... .... . .... ...... ......

... and so on down to lots of files ...

# cd /usr# find . -inum 540488./local/jdbc/ora817/old/libserver8.a

More infos about the RMCmem package can be found here as PDF

Using Sun Solaris Manuals directly from CD-ROM

Solaris 8:

cd /cdrom/sol_8_doc./ab2cd (Start)http://quorum:8888 (Using the Doc online)./ab2cd stop (Stop)

Solaris 7:

cd /cdrom/sol_7_1199_doc./ab2cd (Start)http://diamond:8888 (Using the Doc online)./ab2cd stop (Stop)



DLT-TAPE UNIT INSTALLATION on Solaris 7/8/9

Installation Instructions will cover the installation of the DLT tape peripheral hardware and configuration of thesystem to communicate with the DLT tape peripheral. In this example we use a «QUANTUM DLT7000».

The Solaris system must have the appropriate SCSI interface for DLT drive to attached to, a SCSI single-endedDLT drive can be attached only to a SCSI single-ended interface. The same is true for SCSI differentialattachment. Solaris includes a driver to efficiently communicate with SCSI tape drives, such as the DLT tapeperipheral.

Perform the installation as follows:

Shut down your Sun workstation/server and power off the machine and all scsi-devices. Connect the DLT to thescsi-bus using good cables and make sure the bus is terminated correctly. Set the scsi-id; id 4 or 5 are the mostcommon to use.

If possible use a separate or underutilized SCSI bus for the DLT. Running the tape drive on the same bus as thedisk drives will never let you achieve any good throughput. You bought the DLT because of performance didn'tyou ?

1. STOP-A (L1-A) Power on the devices/machine again and halt the bootprocess with.

(or press the BREAK key if you have an ASCIIconsole).

2. probe-scsi-all Verify that the drive is connected properly.

Note: output from probe-scsi will not always be correct if

you enter the PROM monitor by breaking the boot process!

3. boot -rv Boot the system and log in as root. When booting youshould see a message similar to these: "st1: ".

4. cd /kernel/drv

Change directory to /kernel/drv.

Edit the st.conf file by adding the following:

tape-config-list="QUANTUM DLT7000","Quantum DLT7000","DLT7-data";DLT7-data = 1,0x38,0,0x8639,4,0x82,0x83,0x84,0x85,3;

tape-config-list="<DLT tape unit>","<DLT reference name>","<DLTdata>"

tape-config-list is a variable defined by a series of tape configuration parameters listed below:

<DLT tape unit> is the vendor and product ID string for the DLT device.

Depending on the DLT tape peripheral you are installing, you must insert the appropriate vendor andproduct ID for <DLT tape unit> as described in the following table:

DLT Tape Product <DLT tape unit>

DLT7000 QUANTUM DLT7000 (Total string character

count, including spaces, must equal 15).

<DLT reference name> is a name you select that the system will use to identify the DLT device. This

reference does not change the DLT product ID. When the system boots, the reference name will bedisplayed in the list of peripheral devices recognized by the system.

<DLT-data> is a variable containing a series of additional DLT device configuration information. You select a

name in place of the <DLT data> string. You will continue editing the st.conf file by defining the name you

selected for <DLT data>. The definition depends on the DLT tape peripheral you are installing. For a

DLT7000 series unit add the following line:

1,0x38,0,0x8639,4,0x82,0x83,0x84,0x85,3;

<DLT data> contains 10 parameters and are described following:

1 The first parameter, is the version number and should not change.

0x38 The second parameter, designates the DLT tape type as defined in/usr/include/sys/mtio.h.

#define MT_ISOTHER 0x36 /* generic other type of tape drive */



#define MT_ISOTHER 0x36 /* generic other type of tape drive */#define MT_ISDLT 0x38 /* sun: SCSI DLT tape drive */

0 The third parameter is the block size. Since the DLT tape drive uses variableblock size, this value should be zero.

0x8639 The fourth parameter, 0x8639, is a summation of values that representselected device options. The table below lists the options and thecorresponding value:

Option Value

ST_VARIABLE 0x0001ST_BSF 0x0008ST_BSR 0x0010ST_LONG_ERASE 0x0020ST_NOWS_EOD 0x0200ST_NLOADABLE 0x0400ST_NO_RECSIZE_LIMIT 0x8000

The man st page has more information about these and other possible

device options. For certain applications, it may be necessary to consideradding or removing one or more of the device options.

4 The fifth parameter, 4, defines the number of densities. The maximumdefinable number of densities is 4.

0x820x830x840x85

The sixth, seventh, eighth and ninth parameter are used for systemselection of tape densities. Use these values for a DLT 7000 Tape Drive.

3 The tenth parameter defines which density the system will use as thedefault density. The sixth, seventh, eighth and ninth parameters in the <DLTdata>string are referenced by the system as 0, 1, 2 and 3, respectively.The 3 value for the tenth parameter selects the 0x85 density code as thesystem default density.

After editing the st.conf file, reboot the system:

5. shutdown-i0-g0boot -rv

Reboot the SystemThe -r switch in the boot command enables a kernel compile andincludes the creation of device special files used for communicationwith the DLT device. The -v switch enables verbose mode displayof system bootup. With verbose mode, the system should indicatethat the DLT tape peripheral is attached by displaying the <DLTreference name>string you selected.

6. mt -t /dev/rmt/0 status Enter the following command to verify the installation:

Vendor 'TANDBERG' Product 'DLT7000 ' tape drive:sense key(0x0)= No Additional Sense residual= 0retries= 0 file no= 0 block no= 0

The target drive designations assigned by Solaris may take on values higher than already established in the/dev/rmt/ path. This is not a problem but during a boot -rv, Solaris does not remove tape device files for drives

that are no longer attached to the system. This can increase the effort in locating the device file for theconfigured drive, however, this can be minimized by first deleting the tape device files:

rm /dev/rmt/*

then either boot the system with a:

boot -rv

or issue the following at the command line prompt:

drvconfig -i st; tapes

If the DLTtape is the only drive on the system, it's target assignment should be zero. The Solaris man pages havemore information on drvconfig and tapes.

Reconfigure Devices on Solaris

If you remove or add a device on Solaris then the devices files must be recreated, either with boot -rv or

devfsadm. For example to renumber the logical tape drive devices do the following: Tape drives were numbered

beginning with /dev/rmt/3 instead of /dev/rmt/0. The physical devices pointed to by the logical /dev/rmt/[012]



beginning with /dev/rmt/3 instead of /dev/rmt/0. The physical devices pointed to by the logical /dev/rmt/[012]devices no longer existed, and we wanted to renumber the valid devices beginning at /dev/rmt/0.

1. Cleanup non-existent tape drive devices with devfsadm.

# devfsadm -C -c tape -v

2. Remove all /dev/rmt logical links.

# rm -f /dev/rmt/*

3. Recreate all /dev/rmt logical links with devfsadm

# devfsadm -c tape -v

devfsadm

devfsadm(1M) maintains the /dev and /devices namespaces. It replaces the previous suite of devfs administrationtools including drvconfig(1M), disks(1M), tapes(1M), ports(1M), audlinks(1M), and devlinks(1M).

OPTIONS

The following options are supported:

-C Cleanup mode. Prompt devfsadm to cleanup dangling /dev links that are notnormally removed. If the -c option is also used, devfsadm only cleans up forthe listed devices' classes.

-c device_class Restrict operations to devices of class device_class. Solaris defines thefollowing values for device_class: disk, tape, port, audio, and pseudo.This option may be specified more than once to specify multiple deviceclasses.

OpenBoot Diagnostics

The Solaris operating system gets the jumpstart for its booting from a hardware-level interface called theOpenBoot PROM or OBP for short. OpenBoot at its heart has an interactive command interpreter with a varied setof functions. OBP is a firmware which is stored in the socketed startup PROM of the computer and consists of two

parts, the PROM and the NVRAM.

As stated earlier while the PROM acts as the interface for access to diagnostics and drivers, the NVRAM consistsof some editable user defined parameters. Non Volatile information like the system identification information,device aliases etc are stored in the NVRAM.The OpenBoot PROM is programmable and can be programmed basedon Forth, which is an interactiveprogramming language much like shell scripting.

The main tasks performed by the OpenBoot firmware are:

Initializing and Testing system hardware ( POST , power on self test)Interactive DebuggingManagement of NVRAM ParametersStart the Operating System boot

Useful commands at OK prompt.Dignostics : boot General

banner this command shows the following systems hardware informatiion :Model, architecture,processor,keyboard, openbootversion, Serial no. ethernet address &host id.

test floppy - test floppy disk drivetest net - test network loopbackstest scsi - test scsi interfacetest-all test for all devices withselftest method

watch-clock Show ticks of real-time clock

watch-netMonitor network broadcast packets

watch-net-allMonitor broadcast packets on all netinterfaces

boot - boot kernel from defaultdevice.Factory default is to bootfrom DISK if present, otherwise fromNET.

boot net - boot kernel from networkboot cdrom - boot kernel from CD-ROMboot disk1:h - boot from disk1partition hboot tape - boot default file fromtapeboot disk myunix -as - boot myunixfrom disk with flags "-as"

DEVALIAS

ok>show-devsok cd /pci@1f,4000/scsi@3

ok .properties

ok ls

printenvDisplay all variables and current values. setenv <variable>Set variable to the given value.

set-default <variable>

Reset the value of variable to the factorydefault. set-defaults Reset variable values to the factorydefaults.



interfaces

probe-scsiShow attached SCSI devices probe-scsi-all Show attached SCSI devices for allhost adapters- internal & external.

ok lsf00809d8 tapef007ecdc disk

ok .speed

CPU Speed : 200.00MHzUPA Speed : 100.00MHz PCI Bus A : 66MhzPCI Bus B : 33Mhz

Key SequencesThese commands are disabled if the PROM security is on. Also, if your system has full security enabled, you cannot apply

any of the suggested commands unless you have the password to get to the ok prompt.

Stop - Bypass POST. This command does not depend on security-mode. (Note: some systems bypass POST as a default;

in such cases, use Stop-D to start POST.)

Stop-A Abort.

Stop-D - Enter diagnostic mode (set diag-switch? to true).

Stop-F - Enter Forth on TTYA instead of probing. Use exit to continue with the initialization sequence. Useful if hardware is

broken.

Stop-N Reset NVRAM contents to default values.

Start an OpenBoot Diagnostics

<STOP A>OK setenv diag-switch? trueOK setenv auto-boot? falseOK reset-all

OK test-all or obdiag

Configure Graphics Console (e.g. Sun XVR-100 Graphics Accelerator) instead of serial TTYA

OK show-displaysSelect the graphics accelerator, e.g. b

OK nvalias mydev <CTRL-Y>OK setenv output-device mydevOK setenv use-nvramrc? trueOK reset-all

Why doesn't my .forward file work?

Overview

If you are having problems where you have created a $HOME/.forward file in your home directory to forward e-

mails from one account to another and it just won't forward them?

Set correct Permissions

First make sure the file isn't group or world writable.

-rwxrwxr-x 1 zahn dba 0 Jan 9 12:17 .forward # wrong-rwxr-xr-x 1 zahn dba 0 Jan 9 12:17 .forward # OK

Lastly, make sure your home directory isn't group or world writable.

drwxrwxr-x 14 zahn dba 4096 Jan 9 12:20 zahn # wrongdrwxr-xr-x 14 zahn dba 4096 Jan 9 12:20 zahn # OK

Simple Shell Script to backup your Files

Overview

A backup strategy is more complex than creating a redundant copy of disk storage and consideringthe strategy a success. A successful backup strategy must detail how the backup media are rotated,how the media are archived, how the system will be recovered, and what the backup software will doto create the backup. Although all parts of the backup strategy are equally important, this tip willfocus on the backup script and will detail a flexible backup script that uses built-in Solaris softwaretools which create a reliable local backup of a Solaris machine.

Introduction

The backup script will accomplish the following goals:

Create a backup archive that is as easy to restore a single file as it is to restore an entire filesystem.The backup script will run autonomously. The only human intervention will be to swap media andreview output.



review output.The filesystems or directories to backup can be specified in the script. Using automounter youcan even specify remote filesystems.The script will create a detailed log of the backup.The script will send an abbreviated email summary of the backup to the administrator.After a successful backup, the script will verify to some extent the contents of the backupmedia.The backup script will be able to run on any Solaris 2.6 or greater machine without modification.

Tools used

We use the well known utilities TAR, GZIP and DD, because they are available on any Unix system.They are very well tested and simple to use. In case of an emergency it is important to have a simpleway to restore, independent of complex tools and incompatible software releases.

Magnetic Tape Control

The utility MT sends commands to a tape drive. Many of these commands are familiar, but some are not. The scriptwill use these mt commands.

rewind – rewind the taperewoffl – rewind the tape and eject it (go offline)eom – space to end of recorded media on tapeweof – write count EOF marks at current position on tapestatus – display current status of tape

Script

Click here for the Shell Script

Documents

SUN Solaris Tips