periodic process # do things at right time automatically
cron deamon
wakes up every minute and reads the cron table and executes command on time
/etc/rc.c/cron # FreeBSD
/etc/cron.d # Linux
portaudit
crontab (cron table)
"user's" cron
named the uid
under /var/cron/tabs # FreeBSD
/etc/crontab # system cron
configution # use crontab -e
ignored blank lines or leading spaces and tabs
comments start with '#'
env setting # important
name=value
LOGNAME
SHELL
PATH # full pathname is better
HOME
MAILTO # to record outputs and send to your mailbox
format
min hour day month weekday command
minute 0 .. 59
hour 0 .. 23
day 1 .. 31
month 1 .. 12
weekday 0 .. 6 # 0 for Sunday
matche rules
* everything
- matches ranges
, listed value
/ skips of the number's value through the range
special strings
@reboot # once at startup
@yearly # once a year 0 0 1 1 *
@annually # the same
@monthly # once a month 0 0 1 * *
@weekly # once a week 0 0 * * 0
@daily # once a day 0 0 * * *
@midnight # the same
@hourly # once an hour 0 * * * *
'crontab' command
crontab -e
edit user's cron with your editer such as 'vim'
crontab -l
list the crontab
crontab -r
remove the crontab
crontab filename
install filename as your crontab
to allow and deny users
/var/cron/{allow,deny}
system crontab
/etc/crontab
min hour mday month wday who command
^^^
control the permission as who
periodic
in /etc/crontab
periodic daily
weekly
monthly
/etc/defaults/periodic.conf
/usr/ports/ports-mgmt/portaudit
'at' command
12:15am root@satest # at -t 201110200016
echo Hello world
^D
Job 2 will be executed using /bin/sh
12:15am root@satest # atq
Date Owner Queue Job#
2011年10月20日 周四 00時16分00秒 CST root c 2
'atq'
run jubs queueed by 'at'
'atrun' run per 5 min see /etc/crontab
/var/at/at.{allow,deny}
batch exec commands when system load drops then 1.
===========================================
drivers and kernel
UNIX system
user level programers
kernel
hardware
kernel manages and resourtces
CPU process managemant
memory managemant
I/O devices managemant
kernel provides methods for
IPC inter-process communication
process request resources
system call
user applications
+-- Kernel ---------- | ---------------------+
| syscall interface |
| | | |
| V V |
| filesystem process control |
| | IPC, scheduler |
| device drivers memory managemant |
| | | |
| hardware control |
+------------------- | ---------------------+
hardware
kernel-wide design approaches
monolithic kernels
kernel space, shared memorym fast
Linux, BSD, M$ 9x- ...
microkernels
user space, message passing, slow
Mac OS X, GNU Mach
Hybrid
Monolithic + Micro
M$ NT+
other
Nanokernels
Exokernels
FreeBSD kernel
monolithic -> modular
cannot be dynamical loadable
customized VS GENERIC
fast boot time
lower memory usage
additional hardware support
fine-tune system performance
finding system hardware
dmesg
cat /var/run/dmesg.boot
pciconf -lv
kldstate
kldload
/boot/kernel/*.ko
xx_xxx_load="YES"
kldunload
building FreeBSD source tree
sysinstall
configure -> disrtiburion -> src -> all
customized kernel comfiguration file
sample files
/usr/src/sys/<ARCH>/conf # <ARCH> for i386 or amd64
GENERIC, NOTES, LINT (generated by 'make LINT') <--- to find drivers
/usr/src/sys/conf/NOTES # for arch independent options
customized file
cp GENERIC /usr/local/etc/KERNEL5566
ln -s /usr/local/etc/KERNEL5566 /usr/src/sys/<ARCH>/conf
building a kernal
cd /usr/src
make buildkernel KERNCONF=KERNEL5566
make installkernel KERNCONF=KERNEL5566
make kernel = make buildkernel + make installkernel
kernel locatinos
/boot/kernel/kernel
/boot/kernel -> /boot/kernel.old
when somethings goes wrong
config failed # typo
make failed # fucking errors code 123
# look over your configuration
# or send mails for FreeBSD question mailing
list
boot failed
ok mode # it means your kernel is not ok
boot kernel.old
unload kernel && load/boot/kernel.old && boot
ps failed
recompile and install a word
sysctl command
dynamically set or get kernel parameters
/etc/sysctl.conf
/boot/loader.conf
================================
disk
interface
geometry
add new dicks
install procedure
filesystem check
add a disk using 'sysintall'
RAID
GEOM
disk interface
SCSI # expensive
IDE / ATA
SATA
SAS
USB
disk geometry
secrot
indivifual data block
track
circle
cylinder
circle on all platters
posision
CHS
cylinder, head, sector
disk installation procedure
connecting the disk
IDE : master/slave
SATA
SCSI : ID, terminator
power! # the most important
creating device files
auto created by 'devfs'
formatting
low-level
manufacturer diagnostic utility
partition and lable
allow the disk to be treated as a group of independent data
Ex. root, home, swap partition
suggestion
/var, /tmp => separate partition
establishing logical volumes
compine multiple partitions into a logical volume
software RAID
GEOM
ZFS
'newfs' command
superblock contents # before i-list
inode table's size and location
disk block map
usage information
other filesystem’s parameters
'sync'
mount
Ex.
mount /dev/ad1s1e /home2
automount at boot time
write it in /etc/fstab
swap
swapon, swapoff, swapctl
swapinfo, pstat
'fsck'
system crash will cause
inconsistency between memory image and disk contents
fsck
examine all local filesystem listed in /etc/fstab at boot time.
(fsck -p)
automatically correct the following damages:
unreferenced inodes
inexplicably large link counts
unused data blocks not recorded in block maps
data blocks listed as free but used in file
incorrect summary information in the superblock
RAID # redundant array of inexpensive disks
# combine several physical hard drives into a
logical unit
features
fault tolerance
higher throughput
real-time recovery
level
0~6
hierarchicak RAID (0 + 1)
hardware RAID
software RAID
level
RAID 0
500 + 500 = 1000
stripped data intro several disks
minimum numbers of drivers: 2
advantage
performance
disadvantage
no fault tolerance
RAID 1
500 + 500 = min (500, 500) = 500
minimum numbers of drivers: 2
advantage
100% redundary of data
disadvantage
100% storage overage
slow write performance
Hardware and software
GEOM, ZFS
RAID 2
Hamming Code ECC each bit data word
RAID3 and RAID 4
blocks by blocks
bytes by bytes
parallel transfer with Parity
minimum number of drives: 3
advantages:
very high data transfer rate
disadvantages:
transaction rate equal to that of a single disk drive at best
RAID 5 and RAID 6
independent Disk with distributed parity blocks
minimum number of drives: 3/4
advantage:
highest read data rate
medium write data rate
disadvantage:
disk failure has a medium impact on throughput
complex controller design
when one disk failed, you have to rebuild the RAID array
"write hole" problem
you will see
0, 1, 5, 6
performance, safe, average , more safer than 5
GEOM
ELI
JURNAL
LABEL
MIRROR
SRIPE
major RAID control utilities
kernel modules
/boot/kernel/geom_*
kernel support
{glabel,gmirror,gstripe,g*} load/unload
device GEOM_* in kernel config
geom_*_enable=“YES” in /boot/loader.conf
LABLE
for GEOM provider labelization
kernel
device GEOM_LABEL
geom_label_load=“YES”
'glable' command
# glabel label -v usr da2
# newfs /dev/label/usr
# mount /dev/label/usr /usr
# glabel stop usr
# glabel clear da2
MIRROR # for RAID 1
kernel
device GEOM_MIRROR
geom_mirror_load=“YES”
'gmirror' command
# gmirror label -v -b round-robin data da0
# newfs /dev/mirror/data
# mount /dev/mirror/data /mnt
# gmirror insert data da1
# gmirror forget data <---- this is so funny
# gmirror insert data da1
# gmirror stop data
# gmirror clear da0
SRIPTE # for RAID0
kernel
device GEOM_STRIPE
geom_stripe_load=“YES”
'gstripe' command
# gstripe label -v -s 131072 data da0 da1 da2 da3
# newfs /dev/stripe/data
# mount /dev/stripe/data /mnt
# gstripe stop data
# gstripe clear da0
No comments:
Post a Comment