|
-i
This option is mutually exclusive with all other options except for the -d keyboard
device option. This option instructs the keyboard command to read and process
keyclick and keyboard abort default values from the "Courier New">/etc/default/kbd file. This option can only be used by the
superuser.
-r
off.
abort sequence effect.
Monitoring Performance
More information can be found on "http://docs.sun.com">http://docs.sun.com
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
pagedisk faults cpu
r b w swap free re mf pi po fr de sr f0 s3 -- --in sy cs us sy id
0 0 8 28312 668 0 9 2 01 0 0 0 1 0 0 10 61 82
1 2 97
0 0 3 31940 248 0 10 20 0 26 0 270 4 0 0 53 189 191 6 6 88
0 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 0 121 6 0 0 163 110 138 1 3 96
0 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 52 840 1 99
"border-collapse: collapse" bordercolor="#111111">
Category
Field Name
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 four disks
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 ins0 swap outs
0 pages swapped in0 pages swapped out
409376480 total address trans. faults taken
3075036 page ins2601555 page outs
3812452 pages paged in6525552 pages paged out
11007609 total reclaims
10927650 reclaims from free list0 micro (hat) faults
409376480 minor (as) faults2957386 major faults
102738273 copy-on-write faults
61711047 zero fill page faults1002562077 pages examined by the clock daemon
7881 revolutions of the clock hand16716370 pages freed by the clock daemon
4999048 forks
1138206 vforks5747009 execs
741660225 cpu context switches736047593 device interrupts
528054538 traps
2496638575 system calls
430283487 total name lookups (cache hits 95%)81727 toolong
10484677 user cpu
9528364 system cpu
443762786 idle cpu16281790 wait cpu
How to Display Swapping Statistics
Run vmstat -S to show swapping
statistics.
procs
memorypage
disk
faults cpur b w swap free si so pi po fr de sr m1
m3 m4 m5 in sy cs us sy id0 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 second
so = Number of whole processes swapped out
How to Display Disk Utilization Information (iostat)
You can display disk activity information by using the "Courier New">iostat command with a time interval. The following example
shows disk statistics gathered every five seconds.
iostat 5
"#0000FF">tty
md1
md3
md4md5 cpu
tin tout kps tps serv kps tps serv kps tps serv kps tps
serv us sy wt id0 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
"border-collapse: collapse" bordercolor="#111111">
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.
"#0000FF">
extended device statisticstty cpu
device r/s w/s kr/s kw/s wait actv
svc_t %w %b tin tout us sy wt id
md1 0.4 0.9
3.6 6.9 0.0 0.0 27.7 1
1 0 2 2 2 3 92
md3 0.1 0.2
1.0 1.3 0.0 0.0 21.7 0
0
md4 0.0 0.00.0 0.0 0.0 0.0 0.0
0 0
md5 0.0 0.00.7 0.0 0.0 0.0 9.9
0 0
md8 0.8 0.3 6.714.2 0.0 0.0 13.1 0 1
md10 0.2 0.9 1.8
6.8 0.0 0.0 15.5 0 1
md11 0.2 0.9 1.8
6.8 0.0 0.0 14.8 0 1
md30 0.0 0.2 0.5
1.3 0.0 0.0 11.4 0 0
md31 0.0 0.2 0.5
1.3 0.0 0.0 10.2 0 0
md40 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0 0
md41 0.0 0.0 0.0
0.0 0.0 0.0 0.0 0 0
md50 0.0 0.0 0.4
0.0 0.0 0.0 9.4 0 0
md51 0.0 0.0 0.4
0.0 0.0 0.0 7.3 0 0
md80 0.4 0.3 3.3
14.2 0.0 0.0 10.3 0 0
md81 0.4 0.3 3.3
14.2 0.0 0.0 11.7 0 1
sd0 0.6 2.1 6.0
22.8 0.0 0.0 16.3 0 3
sd1 0.6 2.1 6.0
22.8 0.0 0.0 15.2 0 2
"border-collapse: collapse" bordercolor="#111111">
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 system mode. 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
"border-collapse: collapse" bordercolor="#111111">
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 be enabled 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 metadata changes to the file system are
written. Metadata includes directory and I-node information but not file data blocks,
essentially everything but the actual data within the file. So, for example, a "file
create" modifies the directory structure and allocates a new I-node, and those activities
are written to the logging space. Once the metadata changes are made to the logging area,
the system is free to perform other operations to the file system. In the background, the
information in the log is flushed to the file system and updates the appropriate
directory 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 system to complete metadata changes
via logging and background flushing than by directly modifying the metadata (via random
I/O) spread across the disk. The size of the logging space is based on the size of the
file system, and equals 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 is about to fill up, new metadata change
requests are paused while the log is emptied. As changes are moved from the log to the
file system, that log space is made available, and new metadata changes can be written to
the logging space.
Usually with UFS, if the system crashes during any file
system operation, the entire system must have its consistency checked via the fsck
command. That command can take several minutes per file system because it checks all
metadata and file data to ensure the structures are correct, free, and used, and that the
I-node block counts are correct. It also confirms that the free space available is
current, repairs inconsistencies, and occasionally requires manual intervention to fix
large problems. Files and even directories can be lost, depending on 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 check for a logged file system after a
crash is a simple and fast operation. The system evaluates the logging data and
determines which changes had completed against the underlying file system, which had yet
to start, and which were in progress. Those completed or not yet started are removed from
the log, and those partly completed are either 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 and the activities here. The end
result is that the underlying file system is consistent, and no thorough consistency
checking 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 system configuration file. Logging only
appears in a couple other places within Solaris. The mount command shows which partitions
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 mount
B = Device to fsck
C = Mount point
D = Filesystem Type
E = Fsck pass (unimportatnt with logging)
F = Mount at bootG = Mount options
#
------------------------------------------------------------------
#
A
B
C D EF 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 ufs5 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 ufs8 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 11yes logging
/dev/dsk/c1t14d0s0 /dev/rdsk/c1t14d0s0 /users ufs 12 yes
logging
swap-
/tmp tmpfs - yes -
Logging increases performance, decreases fsck time,
removes the risk of a file system corruption, can be used on all UFS partitions
(including root), and is free.
"Solaris Syslog Daemon Debugging">Solaris Syslog Daemon Debugging
The log system messages daemon syslogd reads and forwards system
messages to the appropriate log files and/or users, depending upon the priority of a
message and the system facility from which it originates. The configuration file
/etc/syslog.conf controls where messages are forwarded. The syslogd daemon ignores
any faulty entry in /etc/syslog.conf, specially spaces instead of tabs are not recognized
by syslogd. Therefore always check the entries in /etc/syslog.conf in the debugging mode
of syslogd.
How to check /etc/syslog.conf
# /etc/init.d/syslog stop
# /usr/sbin/syslogd -d
getnets() found 1 addresses, they are:
0.0.0.0.2.2
amiloghost() 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"
"#0000FF">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 1.70
Started: Sat Jan 6 10:11:47 2001
Input 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/sysmsg
7 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: root
0 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/sysmsg
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 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, operator
X 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 Statistics
FileTot Dups Nofwd Errs
----
--- ---- ----- ----/dev/sysmsg
0 00 0
/var/adm/messages0 0
0 00
0 0 0operator
0 00 0
root0 0
0 0WALL
0 00 0
/dev/sysmsg
0 00 0
/var/adm/messages
0 0
0 0root,operator
0 0
0 0
WALL
0 00 0
syslogd: restarted
off & 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 syslogd will accept the new entry in Line 14.
"Does each Oracle Process use more than 100M memory">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 use more than 100M memory. Is this a
problem on the Oracle installation or something else? It seems that pmap counts the SGA
size as the private memory segment of each oracle process, but we believe the SGA size
should be 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% 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 are actually shared between background processes if
these instances share the same ORACLE_HOME. So you need to subtract 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 text executable 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, we need 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 CMD
oracle 359 1 0 12:26:17
? 0:00 ora_pmon_DIA3oracle 361 1 0 12:26:17
? 0:01 ora_dbw0_DIA3
oracle 363 1 0 12:26:17
? 0:01 ora_lgwr_DIA3oracle 365 1 0 12:26:17
? 0:18 ora_ckpt_DIA3
oracle 367 1 0 12:26:17
? 0:01 ora_smon_DIA3oracle 369 1 0 12:26:17
? 0:00 ora_reco_DIA3
oracle 371 1 0 12:26:17
? 0:00 ora_snp0_DIA3oracle 373 1 0 12:26:17
? 41:50 ora_s000_DIA3
oracle 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"
"#0000FF">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, in this case it is 82992K. So, 124232K minus 82992K
is 41240K. So, the DBWR background process is approximately 41.2 MB. Repeat
this steps for all the background processes.
"Sizing up Solaris Memory with the RMCmem Package">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 of memory and introducing a set of tools, the
RMCmem package.
Install RMCmem Package
Download the RMCmem tools available from "ftp://playground.sun.com/pub/memtool">ftp://playground.sun.com/pub/memtool. The
package includes a kernel module 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 swap space to the system. If you ever see "out of memory"
messages, adding swap space is the usual fix. Performance of the system is very dependent
on how much physical memory (RAM) you have. If you don't have enough RAM to 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
Megabytes
Application:
110 MegabytesExecutable & libs: 42 Megabytes
File
Cache:
757 Megabytes
Free, 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 use the 'pagesize' command to see the size
in bytes per page:
# /usr/sbin/prtconf | grep Memory
Memory size: 1024 Megabytes
# /usr/bin/pagesize
8192
Kernel memory
Kernel memory is allocated to hold the initial kernel code at boot
time, then grows dynamically as new device drivers and kernel modules are used. Kernel
tables also grow dynamically, unlike some older versions of Unix. As you add hardware and
processes to a system, the kernel will grow. In particular, to keep track of all the
memory in 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 grabs memory in large "slabs," then allocates smaller
blocks more efficiently. This means that the kernel tends to grab a bit 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. To show 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 fail
6873088 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 a file, anonymous memory (the swap space),
System V shared memory, or a memory mapped device. The mapped files 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 pmem command in the RMCmem package.
# /opt/RMCmem/bin/pmem 361
361: ora_dbw0_DIA3
Kbytes Resident Shared Private Permissions
Mapped File
82992 8299282992 - read/write/exec [shmid=0x2]
16
16 8 8
read/exec libc_psr.so.116
16 8 8
read/exec libmp.so.2
88 8 -
read/write/exec libmp.so.2........ ..
.. . ............... ...........112
80 72 8
read/exec libelf.so.1
88 8 -
read/write/exec libelf.so.116
16 8 8
read/exec libkvm.so.18
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 currently resident in main memory, wherein 92728 kilobytes
are shared with other processes while 312 kilobytes are private. 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 add in the shared memory for each mapped file, we'll
know how much application memory is in use. This summary is shown 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 92800k808k ora_pmon_DIA3
367 118184k 93152k 92704k
448k ora_smon_DIA3369 117928k 93120k 92704k
416k ora_reco_DIA3
371 118040k 93136k 92720k416k ora_snp0_DIA3
365 119040k 93120k 92712k
408k ora_ckpt_DIA3377 118344k 93080k 92720k
360k ora_d001_DIA3
363 119088k 93056k 92720k336k ora_lgwr_DIA3
375 118344k 93048k 92720k
328k ora_d000_DIA3361 124232k 93040k 92728k
312k ora_dbw0_DIA3
373 121608k 93032k 92728k304k 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 the same 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. The directory 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 the vnode open code) and making an inode-to-name
lookup cache in the kernel. This cache size is limited (to 8192 entries by default), and
the file may have been opened before the kernel module was loaded, so it can't always
find 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 "../download/documents/memtool.pdf">here as PDF
"Using Sun Solaris Manuals directly from CD-ROM">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">DLT-TAPE UNIT INSTALLATION on Solaris
7/8/9
Installation Instructions will cover the
installation of the DLT tape peripheral hardware and configuration of the system 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-ended DLT drive can be attached
only to a SCSI single-ended interface. The same is true for SCSI differential attachment.
Solaris includes a driver to efficiently communicate with SCSI tape drives, such as the
DLT tape peripheral.
Perform the installation as follows:
Shut down your Sun workstation/server and power off the machine and all scsi-devices.
Connect the DLT to the scsi-bus using good cables and make sure the bus is terminated
correctly. Set the scsi-id; id 4 or 5 are the most common to use.
If possible use a separate or underutilized SCSI bus for the DLT. Running the tape
drive on the same bus as the disk drives will never let you achieve any good throughput.
You bought the DLT because of performance didn't you ?
bordercolor="#111111" width="90%" id="AutoNumber1">
1.
STOP-A
(L1-A)
Power on the devices/machine again and halt the boot
process with.
(or press the BREAK key if you have an ASCII
console).
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 you should see a message similar to these: "Courier New" color="#0000FF">"st1: ".
4.
cd /kernel/drv
Change directory to "Courier New">/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 "en-us">tape unit>","<DLT "en-us">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 and product ID for
<DLT tape unit> as described in the
following table:
"#111111" width="80%" id="AutoNumber2">
DLT Tape Product
<DLT tape unit>
DLT7000
Q "en-us">UANTUM 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 be displayed in the list of peripheral devices recognized by the
system.
"Courier New"><DLT-data> "en-us">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 aDLT7000 series unit add the following line:
"Courier New">1,0x38,0,0x8639,4,0x82,0x83,0x84,0x85,3;
<DLT
data> contains 10 parameters and are described
following:
bordercolor="#008000" width="86%" id="AutoNumber3">
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 "#0000FF">/* generic other type of tape drive */
#define MT_ISDLT 0x38 "#0000FF">/* sun: SCSI DLT tape drive */
0
The third parameter
is the block size. Since the DLT tape drive uses variable block
size, this value should be zero.
0x8639
The fourth parameter, 0x8639, is a summation of values that
represent selected device options. The table below lists the options and the
corresponding value:
"en-us">Option
Value
ST_ "en-us">VARIABLE
0x0001
ST_BSF0x0008
ST_BSR
0x0010
ST_ "en-us">LONG_ERASE0x0020
ST_ "en-us">NOWS_EOD0x0200
ST_ "en-us">NLOADABLE0x0400
ST_NO_ "en-us">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 consider adding or removing one or more of
the device options.
4
The fifth parameter, 4, defines the
number of densities. The maximum definable number of densities is 4.
0x82
0x83
0x840x85
The sixth, seventh, eighth and
ninth parameter are used for system selection of tape densities. Use these
values for a DLT 7000 Tape Drive.
3
The tenth parameter
defines which density the system will use as the default
density. The sixth, seventh, eighth and ninth parameters in the <DLT
data>string are referenced by the system as 0, 1, 2 and 3, respectively.
The 3 value for the tenth parameter selects the0x85 density code as the system default density.
After editing the "Courier New">st.conf file, reboot the system:
bordercolor="#111111" width="91%" id="AutoNumber4">
5.
shutdown-i0-g0
boot -rv
Reboot the System
The -r switch in the boot command enables a kernel compile and
includes the creation of device special files used for communication with the DLT
device. The -v switch enables verbose mode display of
system bootup. With verbose mode, the system should indicate that the DLT tape
peripheral is attached by displaying the <DLT reference 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= 0
retries= 0 file no= 0 block no= 0
The target drive designations assigned by Solaris
may take on values higher than already established in the "Courier New">/dev/rmt/ path. This is not a problem but during a "Courier New">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 the configured 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 have more information on
drvconfig and tapes.
"Reconfigure Devices on Solaris">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 "Courier New">devfsadm. For example to renumber the logical tape drive devices
do the following: Tape drives were numbered beginning with "font-style: italic;">/dev/rmt/3 instead of "font-style: italic;">/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 "font-style: italic;">/dev/rmt/0.
- Cleanup non-existent tape drive devices with "font-style: italic;">devfsadm.
# devfsadm -C -c tape -v
- Remove all /dev/rmt logical links.
# rm -f /dev/rmt/*
- 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 administration tools including drvconfig(1M), disks(1M), tapes(1M),
ports(1M), audlinks(1M), and devlinks(1M).
OPTIONS
The following options are supported:
bordercolor="#111111" width="90%" id="AutoNumber5">
-C
Cleanup mode. Prompt devfsadm to cleanup dangling /dev
links that are not normally removed. If the -c option is also used, devfsadm only
cleans up for the listed devices' classes.
-c device_class
Restrict operations to devices of class device_class.
Solaris defines the following values for device_class: disk, tape, port,
audio, and pseudo. This option may be specified more than once to specify
multiple device classes.
"OpenBoot Diagnostics">OpenBoot Diagnostics
The Solaris operating system gets the jumpstart for its booting from a hardware-level
interface called the OpenBoot PROM or OBP for short. OpenBoot at its heart has an
interactive command interpreter with a varied set of 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 consists of 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 based on Forth, which
is an interactive
programming 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 Debugging
- Management of NVRAM Parameters
- Start the Operating System boot
"#EFEFEF" cellpadding="2" cellspacing="0" height="32" width="91%">
Useful commands at OK prompt.
size="1">Dignostics :
boot "1">
General
- banner
- this command shows the following systems
hardware informatiion : Model, architecture, processor,keyboard, openboot
version, Serial no. ethernet address & host id.
test floppy - test floppy disk drive
test net - test network loopbacks
test scsi - test scsi interfacetest-all test for all devices with selftest method
watch-clock
- Show ticks of real-time clock
watch-net
- Monitor network broadcast packets
watch-net-all
- Monitor broadcast packets on all net
interfaces
probe-scsi
- Show attached SCSI devices
- probe-scsi-all
- Show attached SCSI devices for all host
adapters- internal & external.
boot - boot kernel from default device.
Factory default is to boot
from DISK if present, otherwise from NET.
boot net - boot kernel from network
boot cdrom - boot kernel from CD-ROM
boot disk1:h - boot from disk1 partition h
boot tape - boot default file from tapeboot disk myunix -as - boot myunix from disk with flags "-as"
DEVALIAS
- "1">ok>show-devs
- ok cd
/pci@1f,4000/scsi@3 "1">
- ok .properties
- ok size="1">ls
- f00809d8 tape
- f007ecdc disk
- ok size="1">.speed
- CPU Speed : 200.00MHz
- UPA Speed : 100.00MHz
- PCI Bus A : 66Mhz
- PCI Bus B : 33Mhz
- printenv
size="1">Display all variables and current values.
- setenv <variable>
Set variable to the given value.
- set-default "Verdana" size="1"><variable>
- Reset the value of variable to the
factory default.
- set-defaults
- Reset variable values to the factory
defaults.
- Key Sequences
- These 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 "1">- 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 "1"> Abort.
- Stop-D "1">- Enter diagnostic mode (set diag-switch? to
true).
- Stop-F "1">- Enter Forth on TTYA instead of probing. Use exit to continue with
the initialization sequence. Useful if hardware is broken.
- Stop-N size="1">Reset NVRAM contents to default values.
Start an OpenBoot Diagnostics
<STOP A>
OK setenv diag-switch? true
OK
setenv auto-boot? false
OK reset-all
OK test-all or obdiag
Configure Graphics Console (e.g. Sun XVR-100 Graphics Accelerator) instead of serial
TTYA
OK show-displays
Select the graphics accelerator, e.g. b
OK nvalias mydev <CTRL-Y>
OK setenv
output-device mydev
OK setenv use-nvramrc? true
OK
reset-all
Why doesn't my .forward file
work?
Overview
If you are having problems where you have created a "Courier New">$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 "de-ch"> # wrong "#0000FF">
-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 # wrong
drwxr-xr-x 14 zahn dba 4096 Jan 9 12:20 zahn "de-ch"> # OK
Simple Shell Script to backup your Files
Overview
A backup strategy is more complex than creating a redundant copy of
disk storage and considering the 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 do to create the backup. Although all
parts of the backup strategy are equally important, this tip will focus on the backup
script and will detail a flexible backup script that uses built-in Solaris software tools
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 file system.
- The backup script will run autonomously. The only human
intervention will be to swap media and review output.
- The filesystems or directories to backup can be specified in the
script. Using automounter you can 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 backup media.
- 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 simple way to restore,
independent of complex tools and incompatible software releases.
Magnetic Tape
Control
The utility MT size="2">sends commands to a tape drive. Many of these commands are familiar, but some
are not. The script will use these mt commands.
- rewind – rewind the
tape
- rewoffl – rewind the tape and
eject it (go offline)
- eom – space to end of recorded
media on tape
- weof – write count EOF marks
at current position on tape
- status – display current
status of tape
Script
Click "../download/documents/my_backup.txt">here for the Shell Script