is disklist file can contain simple directory ?

2002-06-23 Thread Frederic Saincy 

Hi, 

My (little) problem is the "Can't determine disk and mount point from
$CWD" message during an amrecover session. (restore works if i specify
the _directory_ with "setdisk")

I suppose it comes from my disklist:

panda:~# cat /usr/local/amanda/etc/amanda/DailySet1/disklist
panda.baobab.home/home/olivier  hard-disk-tar

panda:~# cat /etc/fstab | cut -f1-2
# /etc/fstab: static file system information.
#
#  
/dev/hda2   /
/dev/hda1   none
proc
/dev/fd0/floppy
/dev/cdrom  /cdrom

I read in the man that disklist file is suppose to contain mount
points or a disks.

The guess_disk in the debug confirm this, but:

"amadmin DailySet1 panda.baobab.home /home/olivier" 

find the backuped directory.

My questions:

Is there any reason to parse fstab instead of using amadmin parser
wich find disks and directories ?

Is it dangerous to use a directory in disklist file ?

Did i miss something and should not have the "Can't determine disk"
message ? In this case, i can give more information on my
configuration in a next mail. I am using 2.4.3b2.

( no, my laptop's root is not read only, and has not hundred
partitions. Shame on me :-) )

Bye and thank you.

Re: amanda overwrites level0 without warning

2002-06-23 Thread Frederic Saincy 

Hello,

Dave Sherohman <[EMAIL PROTECTED]> writes:

> Sending an advance notice
ok for this.

>> [interactive prompt]
> IMO, no.

I thougth of an interactive prompt with a timeout (after posting to
be honest).

Imagine a scenario like this:

i am an admin or a cron job . In 3 dumps the unique level0 of the
tapecycle will be overwriten.  I do this amdump ( 2 dumps before
overwrite) Admin receive a warning AND a level0 is planned for the
next dump ( 1 dump before overwrite). 

should it be implemented like this ?

I wonder if detecting abnormal operation (with a "max_runs_per_day" in
amanda.conf for example) should be usefull and, if yes, what kind of
action(s) amanda should take.

Anyway, time for me to look at this planner. May the source be with me
:o)

Bye.

Re: amanda overwrites level0 without warning

2002-04-11 Thread Frederic Saincy 

Hello,

Dave Sherohman <[EMAIL PROTECTED]> writes:

> Sending an advance notice
ok for this.

>> [interactive prompt]
> IMO, no.

I thougth of an interactive prompt with a timeout (after posting to
be honest).

Imagine a scenario like this:

i am an admin or a cron job . In 3 dumps the unique level0 of the
tapecycle will be overwriten.  I do this amdump ( 2 dumps before
overwrite) Admin receive a warning AND a level0 is planned for the
next dump ( 1 dump before overwrite). 

should it be implemented like this ?

I wonder if detecting abnormal operation (with a "max_runs_per_day" in
amanda.conf for example) should be usefull and, if yes, what kind of
action(s) amanda should take.

Anyway, time for me to look at this planner. May the source be with me
:o)

Bye.

Re: amanda overwrites level0 without warning

2002-04-11 Thread Frederic Saincy 

Hi,

Was sleepy yesterday. Doing a level0 does not fit, so, for all case,
with or without tape changer / holdingdisk / tape drive (tapeio) what
can be "the" good solution ?

Detect the first abnormal dump, then mail a warning and stop to back
up this tapecycle ?

Does the second dump for a tapecycle in a day can be considered as
"abnormal" ?

Does an interactive prompt for amdump would fit for this case ?

for example:

"Amanda detects an abnormal dumps [big warning] do you *really* know
what your are doing [y/n]?"

Bye.

Re: amanda overwrites level0 without warning

2002-04-10 Thread Frederic Saincy 

Hi,

Frank Smith <[EMAIL PROTECTED]> writes:

> The planner logic might not be set
> up to handle this extreme case.

I don't know very well amanda's internal and this is the occasion for
me to look at the sources. If the planner can't, amdump must wonder:
"If i do this, can admin restore the system ?".

If not, there is 2 solutions in an environment without tape changer:

1) do nothing 
2) do a level0.

I imagine all prefers the second solution.

I gonna look at this "extreme case" further: i did it as a test, but
imagine, for example, a bad written /etc/cronttab and the
The_Big_Problem (r) once level0 was overwritten. Too bad.

Bye.

Re: multiple amdump per day overwrites level0 [long]

2002-04-10 Thread Frederic Saincy 

Hi,

Stephen Carville <[EMAIL PROTECTED]> writes:

>  So, when you hit the seventh amdump, you should have a
> full of every filesystem on one of tapes 2 thru 6 so tape 1 can hold
> partials.

I thougth it works like this too. Amanda simply overwrite the only one
level0 in the set of tapes, without warnings.  let's reiterate the
manipulation.
 
> $ amadmin  find

orion:~$ amadmin DailySet1 find
Scanning /usr/local/amanda/holdingdisk...

date   hostdisk  lv tape or file file status
2002-04-10 orion.progiweb.home /  0 DailySet1-031 OK
2002-04-10 orion.progiweb.home /  1 DailySet1-042 OK
2002-04-10 orion.progiweb.home /  1 DailySet1-052 OK
2002-04-10 orion.progiweb.home /  1 DailySet1-002 OK
2002-04-10 orion.progiweb.home /  1 DailySet1-012 OK
2002-04-10 orion.progiweb.home /  1 DailySet1-021 OK
2002-04-10 orion.progiweb.home /  1 DailySet1-031 OK
2002-04-10 orion.progiweb.home /  0 DailySet1-041 OK
2002-04-10 orion.progiweb.home /home  0 DailySet1-032 OK
2002-04-10 orion.progiweb.home /home  1 DailySet1-041 OK
2002-04-10 orion.progiweb.home /home  1 DailySet1-051 OK
2002-04-10 orion.progiweb.home /home  1 DailySet1-001 OK
2002-04-10 orion.progiweb.home /home  1 DailySet1-011 OK
2002-04-10 orion.progiweb.home /home  1 DailySet1-022 OK
2002-04-10 orion.progiweb.home /home  1 DailySet1-032 OK
2002-04-10 orion.progiweb.home /home  0 DailySet1-042 OK


> $ amadmin  info

Current info for orion.progiweb.home /:
  Stats: dump rates (kps), Full:  655.0, 642.0,  -1.0
Incremental:  947.0, 944.0, 941.0
  compressed size, Full:  37.0%, 37.0%,-100.0%
Incremental:   6.4%,  6.4%,  6.4%
  Dumps: lev datestmp  tape file   origK   compK secs
  0  20020410  DailySet1-041  492800  182109  278

Current info for orion.progiweb.home /home:
  Stats: dump rates (kps), Full:  1006.0, 1003.0,  -1.0
Incremental:  115.0,  -1.0,  -1.0
  compressed size, Full:  61.6%, 61.6%,-100.0%
Incremental:   7.9%,  8.0%,  8.0%
  Dumps: lev datestmp  tape file   origK   compK secs
  0  20020410  DailySet1-042  778770  479897  477

 
> $ amoverview 

 date 04
host disk 10

orion.pr /01
orion.pr /home01


*level0 is on DailySet1-04, i have made it yesterday.

amdump (tape04) l0/l0 (yestday) 
amdump (tape05) l1/l1 (today)
amdump (tape00) l2/l1 (today) (btw why level2 ?)
amdump (tape01) l2/l1 (today)
amdump (tape02) l2/l1 (today)
amdump (tape03) l2/l1 (today)

*Before the overwrite:

> $ amadmin  find

Scanning /usr/local/amanda/holdingdisk...

date   hostdisk  lv tape or file file status
2002-04-10 orion.progiweb.home /  1 DailySet1-042 OK
2002-04-10 orion.progiweb.home /  0 DailySet1-041 OK
2002-04-11 orion.progiweb.home /  1 DailySet1-051 OK
2002-04-11 orion.progiweb.home /  2 DailySet1-002 OK
2002-04-11 orion.progiweb.home /  2 DailySet1-012 OK
2002-04-11 orion.progiweb.home /  2 DailySet1-022 OK
2002-04-11 orion.progiweb.home /  2 DailySet1-032 OK
2002-04-10 orion.progiweb.home /home  1 DailySet1-041 OK
2002-04-10 orion.progiweb.home /home  0 DailySet1-042 OK
2002-04-11 orion.progiweb.home /home  1 DailySet1-052 OK
2002-04-11 orion.progiweb.home /home  1 DailySet1-001 OK
2002-04-11 orion.progiweb.home /home  1 DailySet1-011 OK
2002-04-11 orion.progiweb.home /home  1 DailySet1-021 OK
2002-04-11 orion.progiweb.home /home  1 DailySet1-031 OK

 
> $ amadmin  info

Current info for orion.progiweb.home /:
  Stats: dump rates (kps), Full:  655.0, 642.0,  -1.0
Incremental:  780.0, 776.0, 774.0
  compressed size, Full:  37.0%, 37.0%,-100.0%
Incremental:   4.6%,  4.6%,  4.6%
  Dumps: lev datestmp  tape file   origK   compK secs
  0  20020410  DailySet1-041  492800  182109  278
  1  20020411  DailySet1-051   65710   28604   23
  2  20020411  DailySet1-032   17070 7801

Current info for orion.progiweb.home /home:
  Stats: dump rates (kps), Full:  1006.0, 1003.0,  -1.0
Incremental:  4315.0, 4315.0, 4314.0
  compressed size, Full:  61.6%, 61.6%,-100.0%
Incremental:  66.0%, 66.0%, 66.0%
  Dumps: lev datestmp  tape file   origK   compK secs
  0  20020410  DailySet1-042  778770  479897  477
  1  20020411  DailySet1-031654043151

> $ amoverview 

 date 04 04
host disk 10 11

orion.pr /10 12
orion.pr /home10 11

*After the overwrite:

amdump (tape04) l2/l1 (today)

> $ amadmin  find

Scanning /usr/local/amanda/holdingdisk...

date

multiple amdump per day overwrites level0

2002-04-09 Thread Frederic Saincy 

Hi all,

in my amanda.conf
###
dumpcycle 1 weeks
runspercycle 5
tapecycle 6 tapes
###

i ran 7 amdumps today (testing). The first amdump does a level0, and
the last overwrite this unique level0 with a level1.

This is a bug, a feature or did i miss something ? :)

I can give more information if needed.

Thank you and Bye.

Re: Problems with amrecover

2002-02-26 Thread Frederic Saincy 

Hi!

Joshua Baker-LePain <[EMAIL PROTECTED]> writes:

> Nothing, other than that you probably weren't in the root of a disklist 
> entry when you ran amrecover.  Use 'setdisk' to tell amrecover what disk 
> you want to retrieve files from.

I wonder something:

If disklist entries are directories (not disks) then, you must use
setdisk (even if amrecover were issuing in an disklist entry)

Is it right, or did i miss something ?

I am sorry, i have already post something talking about this:

first post.
http://groups.yahoo.com/group/amanda-users/messages/33524

configuration files & debug.
http://groups.yahoo.com/group/amanda-users/message/33526

Thank you and Bye.

Re: simple directory in disklist file and amrecover

2002-02-25 Thread Frederic Saincy 

Hi,

> Were you issuing the amrecover cmd from /home/oliver?  

>From /home/olivier. Client and server are on the same machine.

I assume it works perfectly for you without using setdisk cmd.

Another problem is that first extract fail, but second works. I think
that i have to rewind the pseudo tape or something like that. I will
look in archives for this, it does not seem to have any relation
between both problems i just mention it in case of.

My configuration files, an amrecover session, and /tmp/amanda/*
follow. ( disklist and fstab files are in previous post )

Revelant (?) extract:

* "Can't determine disk and mount point from $CWD '/home/olivier'" when
issuing a amrecover cmd.

* amindexd.*.debug

> DISK /
< 501 No index records for disk: /. Invalid?
> DISK root
< 501 No index records for disk: root. Invalid?

* amrecover.*.debug

Does a backuped _directory_ is set with guess_disk function ?
How does it look for a similar working config ?


Furthermore, if someone see something weird/useless in this tapeless
configuration feel free to comment :)

Thank you and Bye.




* amanda.conf :

panda:~# grep ^[^#] /usr/local/amanda/etc/amanda/DailySet1/amanda.conf | cut -d "#" -f1
org "DailySet1" 
mailto "backup" 
dumpuser "backup"   
inparallel 4
netusage  600 Kbps  
dumpcycle 1 weeks   
runspercycle 7  
tapecycle 8 tapes   
bumpsize 20 Mb  
bumpdays 1  
bumpmult 4  
etimeout 300
dtimeout 1800   
ctimeout 30 
tapebufs 20 
runtapes 1  
tpchanger "chg-multi"   
changerfile "/usr/local/amanda/etc/amanda/DailySet1/changer.conf"
tapetype HARD-DISK  
labelstr "^DailySet1-[0-9]*$"   
holdingdisk hd1 {
comment "main holding disk"
directory "/usr/local/amanda/holdingdisk"   
use 290 Mb  
chunksize 1Gb   
}
autoflush no
infofile "/usr/local/amanda/etc/amanda/DailySet1/curinfo"   
logdir   "/usr/local/amanda/etc/amanda/DailySet1"   
indexdir "/usr/local/amanda/etc/amanda/DailySet1/index" 
define tapetype HARD-DISK {
 comment "20GB Hard disk"
 length 200 mbytes
}
define dumptype hard-disk-dump {
  comment "Back up to hard disk instead of tape - using dump"
  holdingdisk no
  index yes
  priority high
}
define dumptype hard-disk-tar {
  hard-disk-dump
  comment "Back up to hard disk instead of tape - using tar"
  holdingdisk no
  index yes
  priority high
  program "GNUTAR"
}

* changer.conf :

panda:~# grep ^[^#] /usr/local/amanda/etc/amanda/DailySet1/changer.conf 
multieject 0
gravity 0
needeject 0
ejectdelay 0
statefile /usr/local/amanda/etc/amanda/DailySet1/multi-changer-status
firstslot 0
lastslot 7
slot 0 file:/usr/local/amanda/backup/tape0
[snip 1..6]
slot 7 file:/usr/local/amanda/backup/tape7

* amrecover session :

panda:~# cd /home/olivier/
panda:/home/olivier# hostname -f
panda.baobab.home
panda:/home/olivier# hostname -d
baobab.home
panda:/home/olivier# ls
panda:/home/olivier# rm /tmp/amanda/*
panda:/home/olivier# amrecover 
AMRECOVER Version 2.4.3b2. Contacting server on panda ...
220 panda AMANDA index server (2.4.3b2) ready.
200 Access OK
Setting restore date to today (2002-02-25)
200 Working date set to 2002-02-25.
200 Config set to DailySet1.
200 Dump host set to panda.baobab.home.
Trying disk / ...
Trying disk root ...
Can't determine disk and mount point from $CWD '/home/olivier'
amrecover> setdisk /home/olivier
Warning: no log files found for tape DailySet1-7 written0-00-00
[snip 6-4]
Warning: no log files found for tape DailySet1-3 written0-00-00
Scanning /usr/local/amanda/holdingdisk...
200 Disk set to /home/olivier.
amrecover> settape panda.baobab.home:file:/usr/local/amanda/backup/tape2
Using tape file:/usr/local/amanda/backup/tape2 from server panda.baobab.home.
amrecover> add a*
Added /amanda-2.4.3b2.tar.gz
amrecover> extract

Extracting files using tape drive file:/usr/local/amanda/backup/tape2 on host 
panda.baobab.home.
The following tapes are needed: DailySet1-2

Restoring files into directory /home/olivier
Continue? [Y/n]: 

Load tape DailySet1-2 now
Continue? [Y/n/t]: 
./amanda-2.4.3b2.tar.gz
amrecover> exit
200 Good bye.
panda:/home/olivier# ls
amanda-2.4.3b2.tar.gz
panda:/home/olivier# find /tmp/amanda -type f -name '*debug' -exec echo -e "\n* {}\n" 
\; -exec cat {} \;

* /tmp/amanda/amrecover.20020225091318.debug

amrecover: debug 1 pid 1739 ruid 0 euid 0 start time Mon Feb 25 09:13:18 2002
amrecover: stream_client_privileged: connected to 172.16.16.17.10082
amrecover: stream_client_privileged: our side is 0.0.0.0.745
guess_disk: 13: "/home/olivier"
guess_disk: 0: 1: "/": "/dev/root"
guess_disk: local_disk = 1, fsname = "root"
guess_disk: 1: 5: "/proc": "proc"
guess_disk: 1: 8: "/dev/pts": "devpts"
guess_disk: 1: 13: "/proc/bus/usb": "usbdevfs"
guess_disk: 1: 1: "/": "/dev/hda2"
guess_disk: 1: 5: "/proc": "proc"
g

is disklist file can contain simple directory ?

2002-02-24 Thread Frederic Saincy 

Hi, 

My (little) problem is the "Can't determine disk and mount point from
$CWD" message during an amrecover session. (restore works if i specify
the _directory_ with "setdisk")

I suppose it comes from my disklist:

panda:~# cat /usr/local/amanda/etc/amanda/DailySet1/disklist
panda.baobab.home/home/olivier  hard-disk-tar

panda:~# cat /etc/fstab | cut -f1-2
# /etc/fstab: static file system information.
#
#  
/dev/hda2   /
/dev/hda1   none
proc
/dev/fd0/floppy
/dev/cdrom  /cdrom

I read in the man that disklist file is suppose to contain mount
points or a disks.

The guess_disk in the debug confirm this, but:

"amadmin DailySet1 panda.baobab.home /home/olivier" 

find the backuped directory.

My questions:

Is there any reason to parse fstab instead of using amadmin parser
wich find disks and directories ?

Is it dangerous to use a directory in disklist file ?

Did i miss something and should not have the "Can't determine disk"
message ? In this case, i can give more information on my
configuration in a next mail. I am using 2.4.3b2.

( no, my laptop's root is not read only, and has not hundred
partitions. Shame on me :-) )

Bye and thank you.