Archive for the 'General' Category

Backup & Restore (dump & restore)

Posted by: admin at 1 Monday 29th, 2009

FreeBSD 7.1 and reliable backup procedures published in the test.
The most important is the following.
  • To make a backup of the backup must be able to quickly and reliably is necessary to single-user mode, you need to blank short time.
  • Cumbersome backup process will not be able to continue.
  • To restore OKONAE along the steps, others (including his own after a few months) if you can understand this state must be able to ensure.
I have built a server FreeBSD4.8 period of time the backup was done to a tape device, the recent view that the HDD is a good way to OKONATTA.
Backup to tape, it takes time, and reliability of thin media experience greater capacity than what the HDD is a daily backup tapes must be enormous considering the life of the tape media, real There are無KUNARI.
Once in the past, RAID5 HDD configuration when two cars crashed at the same time, the backup of a few months ago ... I was getting myself in a situation that IMASHI think the best way to get a daily backup However, now we have settled down in this way.

Backup (dump)

Original server and the backup is as follows. 
 ns# df
Filesystem  1K-blocks    Used   Avail Capacity  Mounted on
/dev/ad0s1a   1012974  153114  778824    16%    /
devfs               1       1       0   100%    /dev
/dev/ad0s1f   2025860    9122 1854570     0%    /home
/dev/ad0s1d   6090094 2420298 3182590    43%    /usr
/dev/ad0s1e   1012974   34662  897276     4%    /var

Mount the device to store the backup data

Memory in the USB-in SCSI hard drive, the IDE-HDD to the primary-slave繋GIMASHITA.
Equipment connected to the device name in dmesg to verify. 
 ns# dmesg
.....
ad0: 11264MB <VMware Virtual IDE Hard Drive 00000001> at ata0-master UDMA33
ad1: 13312MB <VMware Virtual IDE Hard Drive 00000001> at ata0-slave UDMA33
.....
見RENAI to scroll faster! If that is 
 ns# dmesg|more
A, [j] or [↓] to scroll at the bottom, [k] or [↑] to scroll up.
[q] to quit the show. (See man more)

Initialize and mount a backup destination

It is necessary only if the first device connected.
Second time is unnecessary. 
 ns# sysinstall
use sysinstall is a very long time, relying on past memories 
 ns# /stand/sysinstall
I, the following error DEMASHITA. 
 /stand/sysinstall: Command not found.
 

 ns# which sysinstall
/usr/sbin/sysinstall
With the recent FreeBSD it is.
Typically, / usr / sbin so should I go to the path, sysinstall is a good reason only.

[Configure] and select [Enter] button.


[Fdisk] select the [Enter] button.


Select the backup destination device [Enter] button. (This is ad1)


[c] and press [Enter] button.


[165] as the [Enter] button.


[w] button.
Confirmation dialog will be displayed [Yes] and select [Enter] button.


[None] Select the [Enter] button.


[q] and press.


[Cancel] and select [Enter] button.


[Label] and select [Enter] button.


[c] and press [Enter] button.


[FS] remains have been selected by the [Enter] button.


Enter the Mount. (The / backup)
Directory is created automatically mount.


[w] button.


[Yes] and select [Enter] button.


[q] menu and press return, [Cancel] button.


[Exit] or [Cancel] button.


[Exit Install] to select [Enter] button.

dump run

Backup of each partition. 
 Filesystem  1K-blocks    Used   Avail Capacity  Mounted on
/dev/ad0s1a   1012974  153114  778824    16%    /           <=== /backup/root.dump
devfs               1       1       0   100%    /dev        <=== バックアップ不要
/dev/ad0s1f   2025860    9122 1854570     0%    /home       <=== /backup/home.dump
/dev/ad0s1d   6090094 2420298 3182590    43%    /usr        <=== /backup/usr.dump
/dev/ad0s1e   1012974   34662  897276     4%    /var        <=== /backup/var.dump
 

 ns# dump -0uaLC 32 -f /backup/root.dump /

Option Description

0 DANPUREBERU 0 (full)
u after a successful dump, / etc / dumpdates database updates.
to detect the end of a media. (Maybe it's no tape.)
L FreeBSD5.0 use the snapshot feature on board.
With this feature, the need to reduce single-user mode.
C 8-32 cache size (unit: MB) by a number between, improving performance dramatically.
However, L is an option if you do not, C is a good option is removed.
(More info)
f you specify a file for backup power.
Likewise, do any of the other three. (High capacity /usr etc. It takes a little time) 
 ns# dump -0uaLC 32 -f /backup/var.dump /var 
 
 ns# dump -0uaLC 32 -f /backup/usr.dump /usr 
 
 ns# dump -0uaLC 32 -f /backup/home.dump /home
This is the complete backup. 

 ns# ls -al /backup
total 2662822
drwxr-xr-x   3 root  wheel            512 Jan 27 17:26 .
drwxr-xr-x  23 root  wheel            512 Jan 27 17:02 ..
drwxrwxr-x   2 root  operator         512 Jan 27 17:02 .snap
-rw-r--r--   1 root  wheel        9646080 Jan 27 17:26 home.dump
-rw-r--r--   1 root  wheel      155197440 Jan 27 17:16 root.dump
-rw-r--r--   1 root  wheel     2524764160 Jan 27 17:36 usr.dump
-rw-r--r--   1 root  wheel       35655680 Jan 27 17:26 var.dump
If you want more safety, the dump file that was created as a temporary file handling, should I copy to another media.

Crunch! For at that moment in the versions of the ftp site to download the livefs.iso LiveFS CD you create.
And "FreeBSD XX emergency boot disk" is written near the server (where glaring) on you.

Restore (restore)

We verify the following configurations.
  • New server (the server boot disk OS What matters is not installed)
  • Above the primary-slave server with the backup data to connect the IDE-HDD
Start HDD ad0
The backup data is stored HDD ad1
Other needs LiveFS CD

Start in LiveFS CD

First, LiveFS CD into the server to start.
(CD does not start if the BIOS is calm Let's review)

If you are good at Japanese, [110 Japan] and select [Enter] button.
[PageUp] 7 times, [↓] 4 is quick and times)

[Japanese 106 keymap] because I have been selected, as [Enter] button.


HDD startup initialization (create partition slice)

[Configure] and select [Enter] button.


[Fdisk] select the [Enter] button.


Select the boot device [Enter] button. (Here is ad0)
Please do not specify a device that is not backed up data is stored.


[c] with the number that appears and press [Enter] button.


[165] as the [Enter] button.


[q] and press.


[Standard] and select [Enter] button.


[Cancel] to return to the menu and press.


[Label] and select [Enter] button.


[c] and then press.


Root (/) to specify the amount assigned to a partition. (Here is 1GB)
Old server (original backup) may have fewer problems than. dd and the problem is not unlike many minutes.


[FS] remains have been selected by the [Enter] button.


Enter the Mount (/), [Enter] button.


Similarly, [c] to set the swap button.
This is the amount of memory that is mounted on the new server is about good times. (Here is 1GB)


[Swap] select the [Enter] button.


Similarly, [c] by pressing the /var to specify the amount to allocate. (Here is 1GB)
Old server (original backup) may have fewer problems than. dd and the problem is not unlike many minutes.


[FS] remains have been selected by the [Enter] button.


Enter the Mount (/ /var button.


Similarly, [c] by pressing the /usr and then assigned to the specified capacity. (Here is 4GB)
Old server (original backup) may have fewer problems than. dd and the problem is not unlike many minutes.
/ home partition and do not create configuration in / usr割RI当TETAI if all the remaining space to the left of the displayed numbers do not specify the amount [Enter] button.


[FS] remains have been selected by the [Enter] button.


Enter the Mount (/ /usr button.


Similarly, [c] all the space left by pressing the / /home assigns.
Do not change the number that appears [Enter] button.
/ home partition is not required if this work is unnecessary.


[FS] remains have been selected by the [Enter] button.


Enter the Mount (/ /home button.


You will like this.
I do not like, select the partition [d] can be deleted by pressing the, please re-set.
If [w] button.


Confirmation dialog will be displayed [Yes] and select [Enter] button.


[q] menu and press return, [Cancel] and select [Enter] button.


Start Fixt

[Fixit] select the [Enter] button.


[CDROM / DVD] to select [Enter] button.


Run fixit.


Device partition you just created (ad0s1a ~ ad0s1f) Make sure you are registered.
Fixit# ls /dev


Partition you just created the / mnt /mnt located below. 
 Fixit# ls /mnt
.snap/  dev/   home/   usr/   var/

Mount the HDD with the backup data

With backup data in order to mount the HDD / mntbk /mntbk a directory. 
 Fixit# mkdir /mntbk
Fixit# _
The IDE-HDD is the Master-Slave connection for the "ad1s1" is the device name.
ad1s1 c of c for, please see here. 
 Fixit# mount /dev/ad1s1c /mntbk
Fixit# _

start restore

ルート（ / ）をrestore 
 Fixit# cd /mnt
Fixit# restore -rf /mntbk/root.dump

/varをrestore 
 Fixit# cd /mnt/var
Fixit# restore -rf /mntbk/var.dump

/usrをrestore 
 Fixit# cd /mnt/usr
Fixit# restore -rf /mntbk/usr.dump

/homeをrestore 
 Fixit# cd /mnt/home
Fixit# restore -rf /mntbk/home.dump

In many cases, usr and restore a large error in the following. 
 Fixit# cd /mnt/usr
Fixit# restore -rf /mntbk/usr.dump

/: write failed, filesystem is full
restore: /tmp//rstdir1233044785: cannot write directory database: No space left on device
Try making space in /tmp/, or set environment variable TMPDIR
to an alternate location with more disk space.
Fixit# _
This is, restore occurs because the amount of work required for temporary files.
If so, the following method to change the TMP again please restore directory. 
 Fixit# mkdir /mntbk/tmp
Fixit# TMPDIR=/mntbk/tmp
Fixit# export TMPDIR
/ mntbk the HDD is a matter of data backup.
TMP here to create a directory means that you use.
After this, the need is some SEMARA / mntbk / tmp if you delete a directory will prevent almost all commands are executed.
In this case, as follows: Replace the TMP directory. 
 Fixit# TMPDIR=/tmp
Fixit# export TMPDIR

Rewriting fstab

It is not necessary to rewrite work, let's check.
The difference between the order of the partition created, led to a situation that does not start, I caught two days ...

Make sure to mount the current situation. 
 Fixit# df
Filesystem      1K-blocks    Used   Avail Capacity  Mounted on
/dev/md0             4175    2449    1726    59%    /
devfs                   1       1       0   100%    /dev
/dev/ad0s1a       1012974  153114  778524    16%    /mnt
devfs                   1       1       0   100%    /mnt/dev
/mnt/dev/ad0s1f   8122034    9426 7462846     0%    /mnt/home
 /mnt/dev/ad0s1e   4058062 2434140 1299278    65%    /mnt/usr 
 /mnt/dev/ad0s1d   1012974   34750  897180     4%    /mnt/var 
/dev/acd0          226150  226150       0   100%    /dist
/dev/ad1s1c      13198878 2662822 9480146    22%    /mntbk

Then open your fstab.
Fixit# vi /mnt/etc/fstab 
 # Device          Mountpoint         FStype   Options         Dump     Pass#
/dev/ad0s1b       none               swap     sw              0        0
/dev/ad0s1a       /                  ufs      rw              1        1
/dev/ad0s1f       /home              ufs      rw              2        2
 /dev/ad0s1d       /usr                ufs      rw              2        2
 /dev/ad0s1e       /var                ufs      rw              2        2
/dev/acd0         /cdrom             cd9660   ro,noauto       0        0

I notice the differences underlined.
Device name "ad0s1e" and "ad0s1d" have not been mounted in the reverse name.
If you like this, "/ mnt / etc / fstab" opens in the editor, please save corrected.

That's it!
LiveFS CD without a rise to restart.

Aftertreatment

When you perform a restore, the current directory of each partition will create a file called restoresymtable.
This is what will be required to restore an incremental backup, restore and remove you when you are finished. 
 ns# ls /
.cshrc        boot          entropy          mnt             sys
.profile      cdrom         etc              proc            tmp
.snap         compat        home             rescue          usr
COPYRIGHT     dev           lib               restoresymtable  var
backup        disk2         libexec          root
bin           dist          media            sbin
 

 ns# rm /restoresymtable 
 ns# rm /var/restoresymtable 
 ns# rm /usr/restoresymtable 
 ns# rm /home/restoresymtable

At this

Long stop in the following message when you try to start the restore server. 
 Configuring syscons: keymap blanktime
"Configuring syscons: keymap blanktime" Please refer to the waiting time out of view.

Can not partition out the following message: 
 Unable to find device node for /dev/ad0s1b in /dev!
The creation of filesystems will be aborted.
Fdisk to create the slices in the [w] Do not press?
Fdisk menu [w] is押SAZU, [q] and press only, Label from the end of the set, the last in the Label menu [w] button.

fails in sysinstall Fdisk Label
Have experienced many times during the verification.
Fdisk Label in sysinstall please refer to fail.

Want to backup another remote server that is connected to the network
The old story, "the tape drive to backup & restore" Please help.

Filed under install, General |

1 Comment »

"Configuring syscons: keymap blanktime" timeout waiting for display

Posted by: admin at 1 Monday 27th, 2009 

 Configuring syscons: keymap blanktime
Have shown that the stop time is caused by the sendmail start Then you're waiting for the timeout to not start correctly.
DNS, and network around見直SHIMASHOU.

If I do, dump & restore of VMware testing encounter this problem while on the go, because he was the wrong network adapter settings ...

Filed under General |

1 Comment »

sshd configuration (DSA authentication)

Posted by: admin at 11 Monday 8th, 2008

FreeBSD sshd is by default to boot. 
 # which sshd
/usr/sbin/sshd
So, after the /etc/ssh/sshd_config to AllowUsers USERNAME only that, BURUTOFOSUATAKKU (brute force attacks) against the DSA should be authenticated.
If you look you see the access log, crackers or a humble country or分RIMASEN bots, you should find that traces of the attack.

Rewrite of the configuration file

/etc/ssh/sshd_config 
 Protocol 2
SSH2 protocol used 
 PermitRootLogin no
disable root login in 
 PubkeyAuthentication yes
Authentication Using DSA 
 PasswordAuthentication no
PermitEmptyPasswords no
No and no and no password authentication and password 
 ChallengeResponseAuthentication no
So no to no the password on it.

After rewriting sshd restart. 
 # /etc/rc.d/sshd restart
Or, in earlier versions may also be an error because, if it is in the following way. 
 # ps aux | grep sshd
# kill -HUP  PID

Private key id_dsa and public key id_dsa.pub to generate

Log in with ssh to enable users to access. root other)
Enter the following command. 
 $ ssh-keygen -d 
 Generating public/private dsa key pair.
Enter file in which to save the key (/home/ USERNAME /.ssh/id_dsa):    [Enter] 
Created directory '/home/ USERNAME /.ssh'.
Enter passphrase (empty for no passphrase):                      [PASSWORD] [Enter] 
Enter same passphrase again:                                     [Re PASSWORD] [Enter] 
Your identification has been saved in /home/ USERNAME /.ssh/id_dsa.
Your public key has been saved in /home/ USERNAME /.ssh/id_dsa.pub.
The key fingerprint is:
xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx:xx  USERNAME @ HOST . DOMAIN . NAME

Public key id_dsa.pub to rename the 

 $ cd /home/ USERNAME /.ssh
$ mv id_dsa.pub authorized_keys
$ chmod 0600 authorized_keys

Private key id_dsa to move to a client machine

ssh client to connect to the /home/ USERNAME /.ssh/id_dsa move. If you are concerned about security, FTP, etc. There is no USB floppy disk is a good memory and do in the process of using the media attention.
id_dsa to save the file, it will depend on the client terminal, please refer to your manual.
Personally open source Tera Term is a good idea.

Filed under General |

No Comments »

To a specific user root grant

Posted by: admin at 11 Monday 3rd, 2008

At the command line user su and type is set to be able to gain root privileges.
/etc/group 
 wheel:*:0:root, USERNAME

Filed under General |

No Comments »

To stop and restart a specific service

Posted by: admin at 11 Monday 3rd, 2008

Stop 
 # kill -KILL `cat /var/run/ PROCESS .pid`

Restart 
 # kill -HUP `cat /var/run/ PROCESS .pid`

Check the process ID or the ID specified in the process.
Stop 
 # ps -aux | grep named
# kill -KILL　 PID
Restart 
 # ps -aux | grep named
# kill -HUP　 PID

Filed under General |

No Comments »

Process ID examine the

Posted by: admin at 11 Monday 3rd, 2008 

 # ps -aux | grep  PROCESS_NAME

Filed under General |

No Comments »