Re: amanda.conf w/ qualstar tape library question

[John Koenig]

What I am having some problems with is the amanda.conf section that deals with
tpchanger and such.  My robot is conrtoled with the device name of 
/dev/sg/h0c0t010

Here is a ( porrly done and not working ) section of my config file:

runtapes 1  # number of tapes to be used in a single run of amdump
tpchanger "chg-zd-mtx"  # the tape-changer glue script
tapedev "/dev/nst0" # the no-rewind tape device to be used
rawtapedev "/dev/nst0"  # the raw device to be used (ftape only)
changerfile "/usr/share/doc/mtx-1.2.16/contrib/mtx-changer"
#changerfile "/var/lib/amanda/DailySet1/changer-status"
#changerfile "/etc/amanda/DailySet1/changer.conf"
changerdev "/dev/nst0"


changerdev might be one thing to look at...
You're so close...

changerdev /dev/sg/h0c0t010

Re: chg-zd-mtx-2.4.3b4 problem

[John Koenig]

At 1:09 PM -0600 10/29/02, Paul T. Root wrote:

Hi,
	I have a problem with the zd-mtx script. I traced it down
to the point in the script where it decides it needs to run the
cleaning tape. Line 968 or so:




I do not have a useful answer. Instead, another question.
This cleaning stuff is new to my radar screen...

Is the "chg-zd-mtx" changer module that exists between AMANDA and the 
MTX program supposed to determine when a cleaning occurs? (or any 
changer module for that matter?)

Re: What features should I look for in a new library?

[John Koenig]

Maybe if I can help by formulating a question.

(I am not addressing your other concerns)



Multiple tape drives:
Our old library had only one drive installed.  I could have added
a second one, but I could not see how amanda would use it to advantage.
Can amanda use a second (or third...) drive?
Is there any benefit?


I think the end-game concern should be related to software support

Maybe the question needs to be asked if anyone has successfully used 
an existing AMANDA changer module with two tape drives in a single 
media changer and in a single AMANDA changer configuration... If 
so... I want to know about it...

I would presume (hopefully I am wrong) that people who have done this 
modified their changer module code locally and it has not been rolled 
into the AMANDA source.

Re: questions about proposed amanda hardware solution

[John Koenig]

>The amanda control host would be a smallish Linux rackmount job with
>attached SCSI+RAID disk enclosure -- currently I'm considering a Dell
>PowerEdge 1650 (1U, PIII-based) with a Dell PowerVault 220S (3U, 14 x
>36 GB Ultra160 SCSI disks + PERC/3 RAID controller).  Doing a RAID-5 +
>hot spare across the fourteen disks would give me 400 "real" GB
>(1024^3) of holding disk.


One comment...

When I was pricing holding disk storage about 6 months ago, I came to 
the conclusion that I would rather pay a buck or $1.25 per GB (ATA 
100) than the $6-$6.50 per GB premium for Ultra 160...

Though this may alter your vertical height requirements as I am not 
aware of any high-end rackmount IDE/ATA storage solutions...but I 
presume they exist... as it would be a business opportunity due to 
the cost per GB.

Re: reply-to in mailing list

[John Koenig]

>  >   o A stronger community results when individuals respond to
>>  individuals as opposed to responding to all members of the the list...
>>
>
>I don't doubt that you are right about that, but I do read threads that have
>nothing to do with me.  And I have gained quite a bit of personal knowledge
>by doing so.  So when I have a problem and ask the list for help, I like to
>think that by replying to the list I help to further other's knowledge as
>well... but maybe thats just me ;)


Yes.. that is the ideal  :) ... but take a look at what happens to a 
list when reply-to is set to the list and the people do not exercise 
*a lot* of self-discipline...

I just recently un-subscribed to just such a list because the traffic 
and noise just became intolerable...  And I was REALLY interested in 
the subject matter too! So it was not as easy decision... at first. 
The list was an otherwise useful list except that personal attacks 
and me too's, and jsut and overall chattiness prevailed... It was all 
posted to the list... I finally through in the towel. In contrast, I 
cannot (do not want to) afford to leave the AMANDA list any time soon.

Plus, there are some *way* smart people here (with more experience in 
their little finger than...). The community risks losing these people 
if their busy lives (or personalities) cannot tolerate excessive list 
traffic. In my opinion, the daily traffic is high enough as 
configured...

:)

Re: reply-to in mailing list

[John Koenig]

how about parsing for anything in the Sender header

"Sender: [EMAIL PROTECTED]"

and be done with it...


###

And regarding the reply-to to the list...

I think (hope) not.   :)


I think posting to the list should be a little extra effort as this 
is not a small community.


  o Traffic would increase with no significant increase in content.

  o If a response needs to go to the list because the information 
needs to be documented, then it will probably be evident by the 
nature of the specific thread.

and last but not least:

  o A stronger community results when individuals respond to 
individuals as opposed to responding to all members of the the list...

Re: New tape nonsense

[John Koenig]

Related question

>
>
>A single amdump run if larger than a single tape can be placed on
>multiple tapes if the runtapes parameter is > 1.


I have a scenario with a changer wherein the runtapes is 1 however 
some days it does not manage to schedule the backups such that it 
uses only one tape...

Are there some guidelines one can use to modify  parameters in 
amanda.conf such that the behavior of AMANDA changes enough to use 
one tape per run instead of one tape 70% of the time?

Or should I post a representative history of my amdump summaries with 
the hope that someone can make specific suggestions, that way?

Re: ERROR: no tape online

[John Koenig]

>
>OTOH, I cheerfully build and install every snapshot Jean-Louis puts
>up on his site, and a lot has been "smoothed up" in the current
>2.4.3b4 snapshots.  I might be worth the effort, maybe even the
>changer stuff works for you now, its running mine just fine.


I am thinking my issue is not related to a particular version of 
AMANDA. Besides, I cannot try 3b3 of any build considering the fact 
that it will not build on the Tru64 platform. So I am stuck with the 
tools at hand.


*** I would think that my media changer should not be incurring
*** an I/O error after the tape was loaded and after a successful
*** amcheck confirmation.

Does anyone know what would happen if this configuration was that of 
a changer? How would AMANDA/amdump/amtape proceed after that error?

Moreover, how is the changer module supposed to handle this I/O 
error? What should the changer pass back to amtape on stdout/stderr?


Thanks in advance.. I hope some folks can provide some guidance and 
more suggestions...

I am attempting to complete the development of a changer module for 
this platform, so useful information, for these purposes, would be 
very helpful...

thx!
j

...

>
>>  ### ### ###
>>
>>Initially I ran the amcheck to remind myself of what tape was
>>  next...
>>
>>$ amcheck DailySet2
>>Amanda Tape Server Host Check
>>-
>>Holding disk /net/redeye/holding/amanda: 177292000 KB disk space
>>  available, that's plenty ERROR: /dev/ntape/tape0: no tape online
>>(expecting tape DailySet2-10 or a new tape)
>>NOTE: skipping tape-writable test
>>Server check took 119.557 seconds
>>
>>  ### ### ###
>>
>>
>>  Cool.. that's slot 9!
>>
>>$ robot load slot 9
>>LOADING SLOT 9 into DRIVE 0 using ROBOT /dev/changer/mc0.
>>
>>
>>  ### ### ###
>>
>>
>>$ nohup amdump DailySet2 &
>>
>>  output of amdump sent via e-mail and presented below...
>>
>>
>>  ### ### ###
>>
>>$ robot unload slot 9
>>ROBOT /dev/changer/mc0 is not responding: I/O error.
>>
>>
>>  H... not good...
>>  ${?} = 64 but Compaq does not
>>  provide the damn info. F*ckers.
>>
>>
>>  ### ### ###
>>
>>$ robot unload slot 9
>>UNLOAD DRIVE 0 into SLOT 9 using ROBOT /dev/changer/mc0.
>>
>>
>>  tape successfully unloaded
>>
>>=
>>
>>Output of amdump:
>>
>>
>>*** A TAPE ERROR OCCURRED: [no tape online].
>>Some dumps may have been left in the holding disk.
>>Run amflush to flush them to tape.
>>The next tape Amanda expects to use is: DailySet2-10.
>  >
>>

Re: Amanda 2.4.2p2 won't backup the backup server...

[John Koenig]

>
>Here's the /etc/inetd.conf lines relating to amanda:
>amandaidx   stream  tcp nowait  root.root 
>/usr/local/inst/backup/amanda-2.4.2p2/libexec/amindexd amindexd
>amidxtape   stream  tcp nowait  root.root 
>/usr/local/inst/backup/amanda-2.4.2p2/libexec/amidxtaped 
>amidxtaped
>amanda dgram   udp waitroot.root 
>/usr/local/inst/backup/amanda-2.4.2p2/libexec/amandad   amandad
>


here's a quick guess...
maybe this needs to be the amanda user?

J

ERROR: no tape online

[John Koenig]

I need some suggestions from the community... I am trying to debug a
problem however I am not sure how/where to look for more clues. 
I am open to any and all comments anyone has time to express.
... and thanks!!

The issue is when amdump is run, something weird happens
with the tape causing AMANDA to go into degraded mode... leaving the
backups on the holding disk.

I have an AIT2 tape media robot (Compaq SSL2020) driven by
the software called Media Robot Utility v1.5

I have been performing the role of the changer glue manually whenever I
run actual backups... The AMANDA configuration is not configured as a 
tape changer at present. 

This last run had a problem... Here is some information about it:

### ### ###

Initially I ran the amcheck to remind myself of what tape was next...

$ amcheck DailySet2
Amanda Tape Server Host Check
-
Holding disk /net/redeye/holding/amanda: 177292000 KB disk space available, that's 
plenty
ERROR: /dev/ntape/tape0: no tape online
   (expecting tape DailySet2-10 or a new tape)
NOTE: skipping tape-writable test
Server check took 119.557 seconds

### ### ###


Cool.. that's slot 9!

$ robot load slot 9
LOADING SLOT 9 into DRIVE 0 using ROBOT /dev/changer/mc0.


### ### ###


$ nohup amdump DailySet2 &

output of amdump sent via e-mail and presented below...


### ### ###

$ robot unload slot 9
ROBOT /dev/changer/mc0 is not responding: I/O error.  


H... not good...
${?} = 64 but Compaq does not 
provide the damn info. F*ckers.


### ### ###

$ robot unload slot 9
UNLOAD DRIVE 0 into SLOT 9 using ROBOT /dev/changer/mc0.


tape successfully unloaded

=

Output of amdump:


*** A TAPE ERROR OCCURRED: [no tape online].
Some dumps may have been left in the holding disk.
Run amflush to flush them to tape.
The next tape Amanda expects to use is: DailySet2-10.


STATISTICS:
  Total   Full  Daily
      
Estimate Time (hrs:min)0:25
Run Time (hrs:min) 0:33
Dump Time (hrs:min)0:08   0:00   0:08
Output Size (meg) 447.30.0  447.3
Original Size (meg)   832.60.0  832.6
Avg Compressed Size (%)53.7--53.7   (level:#disks ...)
Filesystems Dumped   19  0 19   (1:18 2:1)
Avg Dump Rate (k/s)   975.5--   975.5

Tape Time (hrs:min)0:00   0:00   0:00
Tape Size (meg) 0.00.00.0
Tape Used (%)   0.00.00.0
Filesystems Taped 0  0  0
Avg Tp Write Rate (k/s) -- -- -- 


NOTES:
  planner: Incremental of redeye:/net/adder/scratch1 bumped to level 2.
  planner: Full dump of redeye:/net/adder/scratch2 promoted from 11 days ahead.
  planner: Full dump of bum:/net/bum/srv promoted from 11 days ahead.
  planner: Full dump of redeye:/var/amanda promoted from 11 days ahead.
  planner: Full dump of redeye:/net/redeye/home1 promoted from 11 days ahead.
  planner: Full dump of phish:/usr/local promoted from 11 days ahead.
  planner: Full dump of redeye:/net/redeye/lapb1 promoted from 11 days ahead.
  planner: Full dump of phish:/home/phish promoted from 11 days ahead.
  planner: Full dump of ringley:/net/ringley/camb1 promoted from 11 days ahead.
  planner: Full dump of ringley:/net/ringley/tamu1 promoted from 11 days ahead.
  planner: Full dump of redeye:/etc promoted from 11 days ahead.
  planner: Full dump of ringley:/etc promoted from 11 days ahead.
  planner: Full dump of bum:/etc promoted from 11 days ahead.
  planner: Full dump of anabelle:/etc promoted from 11 days ahead.
  planner: Full dump of redeye:/usr/local_redeye/etc promoted from 11 days ahead.
  planner: Full dump of phish:/etc promoted from 11 days ahead.
  planner: Full dump of bum:/var/yp promoted from 11 days ahead.


DUMP SUMMARY:
 DUMPER STATS  
   TAPER STATS   
HOSTNAME   DISK  LORIG-KB OUT-KB   COMP% MMM:SS
 KB/s   MMM:SS KB/s
-- 
- -
anabelle  /etc   1 60 3253.3   0:01
 59.4 N/A  N/A 
anabelle  /net/anabelle/scratch1 1 174750 16630495.2   1:15   
2211.5 N/A  N/A 
bum   /etc   1 50 3264.0   0:00
120.0 N/A  N/A 
bum   /net/bum/srv   1  25900   342413.2   0:10
351.1 N/A  N/A 
bum   /var/yp1 10 32   320.0   0:00
300.6 N/A  N/A 
phish /etc   1   

tape error

[John Koenig]

Anyone seen this tape error message before?

Does this indicate a bad tape or a tape wherein the AMANDA label has 
been inadvertently removed?

thanks,
J


###


Subject: DailySet2 AMFLUSH MAIL REPORT FOR September 13, 2002

*** A TAPE ERROR OCCURRED: [reading label: No space left on device].
 ^^



-- 

*** Qmail-Scanner Envelope Details Begin ***
X-Qmail-Scanner-Mail-From: "[EMAIL PROTECTED]" via maat.reeusda.gov
X-Qmail-Scanner-Rcpt-To: "[EMAIL PROTECTED]"
X-Qmail-Scanner: 1.12 (hbedv: 2.0.3/vdf=6.14.0.3  Clear:. Processed in 0.267526 secs)
*** Qmail-Scanner Envelope Details End ***

Re: please explain output

[John Koenig]

It is probably you font settings... try again but with larger characters.
actually  :)
You did not specify your CWD when running this command...
what is your cwd when you attempt to run the command?
I believe it needs to be your ./DailySet1 dir or whatever it is called...

At 1:14 PM -0700 8/30/02, Chris Bourne wrote:
>  Can someone plaese explain to me the output of this command???
>
>bash-2.05$ /usr/lib/amanda/chg-scsi -info
>chg-scsi: open: /dev/sg0: Success
> open: /dev/sg0: Success
>


-- 

Re: chg-zd-mtx on freebsd

[John Koenig]

I would recommend not trying  to debug the changer stuff from way up 
on high in AMANDA, but to debug at the command line by running the 
changer script directly... The ./docs/TAPE.CHANGERS document will 
guide you with the syntax to do this on the CL.

If you have not already made all of AMANDA operate without the 
changer (just set the tapedev to the device specifier), then I would 
recommend doing that first

Isolating the two will help the issues fall out in a more apparent fashion...


-- 

Re: could not get changer info

[John Koenig]

>I have tried other scripts and have had no success but this is the one I
>have put the most time into because I know the machine uses scsi commands


the amanda changer is middle-ware between the amanda software such as 
amtape and the media changing software.

Therefore one's decision is not based upon transport but instead 
based upon which amanda glue code is designed to handle the 
communication protocol with the media changer which is abstracting 
those details of the transport and the OS's interface make it go.

HTH

-- 

Re: Run and dump cycle recommendations?

[John Koenig]

>
>After seeing that /dev/sg0 is a rewinding device and reading in INSTALL
>that it needs to point to a non-rewinding device, I changed it to point
>to /dev/nst0 which is supposedly it in linux.  However, when 'amcheck -s
>DailySet' is run again, I get this:
>
>---
>Holding disk /backup/amanda: 12891648 KB disk space available, using
>7771648 KB
>ERROR: /dev/nst0: rewinding tape: No medium found
>(expecting a new tape)
>NOTE: skipping tape-writable test
>NOTE: info dir /var/lib/amanda/DailySet1/curinfo: does not exist
>NOTE: it will be created on the next run
>NOTE: index dir /var/lib/amanda/DailySet1/index/localhost: does not
>exist
>Server check took 30.246 seconds
>---
>
>At least with /dev/sg0, it had found a medium.  And I'm not too sure
>what's going on with the permission denied error either.



Was there a tape inserted into the drive at this point?
Had it been amlabel'ed

hopefully that is what is missing at this juncture...

?


-- 

Re: chg-zd-mtx output

[John Koenig]

>On Thu, Aug 15, 2002 at 12:49:04PM -0400, Jason Greenberg wrote:
>>  Does anyone know what could cause this output?  I am trying to debug my
>>  setup with a PowerVault 128T / Linux Redhat 7.3
>>
>>
>>  bash-2.05a$ /usr/lib/amanda/chg-zd-mtx -info
>>  /usr/lib/amanda/chg-zd-mtx: [: : integer expression expected
>>  /usr/lib/amanda/chg-zd-mtx: [: -lt: unary operator expected
>>16 1 1


I'll take a chance of clarifying instead of muddying this issue
 From what I have been learning from some FINE people on this list 
about shell scripting is... that the error above is likely caused by 
a string of length zero as one of the variables in the test. If the 
source was written with by prepending 'x' on both variables such that 
the value of the variables would be concatenated with the 'x', the 
broken test would not likely occur. I think this explanation is 
accurate but it needs review.


>
>where the script is blowing up, but in the mildly amazing line like:
>
> usedslot=`echo $tmpslot |
>   sed -n "s/Data Transfer Element $drivenum:Empty/-1/p;
> s/Data Transfer Element $drivenum:Full (Storage 
>Element \([1-9][0-9]*\) Loaded)\(.*\)/\1/p"`


Now, broken-ness at this point may indicate the media changer's 
output is not compatible with the chosen amtape changer glue code 
module...



>
>where the sed is not interacting w/the mtx output from your changer the
>way the script expects.

yes



-- 

Re: Holding Disk Question

[John Koenig]

>My motto:  Disk is cheap, don't skimp on holding disk.
>

Yup... after I deployed a 181 GB drive for our holding disk, I saw 
somewhat different behavior in amstatus... though I would have 
expected the total time of the backups to drop (compared to a 
relatively small 25 GB holding space I used before) they did not... I 
conclude this was because the entire run is tape i/o bound... Holding 
disk usage on the runs with this disk were about 90% capacity... so 
the clients were not taking so long to complete their data 
transfers...

This would seem to indicate I need to double this amount of space to 
hold 2 days of backups, currently... I'd like to have a week's worth, 
actually in case the tape goes South and replacement is not easy 
and quick.

So does the chunksize parameter affect performance in any way?

thx


-- 

Re: exclude files

[John Koenig]

>please help,
>
>can you exclude files using regular dump or do you
>have to use gnutar?
>

I have a question about this...

What is the relationship of the entry related to exclude files in the 
dumptype root-tar (and its "descendants") and the configure option 
--with-gnutar-listdir=path ???   What pitfalls might happen on the 
client, depending on the configure options that the client was built 
with...?

I wonder if there is no relationship and I was chasing my tail? Is it 
handy to be able to specify this via a dumptypes? Must it be 
specified in the client configuration before building the programs?

I experienced a lot of mucking about on this issue whereby I was 
chasing down errors related to exclude files on a particular Solaris 
client... It was awful and persisted for an hour causing body aches 
and fever. :)  Of course it is possible I was chasing my tail however 
I thought that the previous build was not working due to amcheck 
errors spewed by the tape server regarding a ill-defined exclude 
list...

Anyone have a view, from a vista higher than mine, on this topic?

thx
J
-- 

Re: Tapeless error with 2.4.3 (expecting a new tape)

[John Koenig]

>Hi John,
>
>Thanks for the response. I have also set it up the way that you are talking
>about but I still have the same problem. Could it be the directory names
>that you are using? What is your labelstr entry in your amanda.conf file set
>to?


You may need to label your tape, next.

This would not be related to labelstr, unless of course you are 
indeed doing the label step and the literal you are using does not 
match the pattern in the labelstr.



-- 

Re: Tapeless error with 2.4.3 (expecting a new tape)

[John Koenig]

>I have amanda set up for a tapeless backup. My tapedev is set to the 
>following:
> tapedev "/usr/local/amanda/backup"
>
>In the /usr/local/amanda/backup directory, there is a data directory.
>
>When I run amcheck backup, I get the following error:
>file:/usr/local/amanda/backup : not an amanda tape (expecting a new tape)
>
>

based on your current configuration stated above:

Make as many directories as you desire inside "/usr/local/amanda/backup"

then use a softlink named data to point to the directory that is the 
next "tape"...

I created 10 dirs named "slot_00, slot_01...slot_09" then I rotate 
the softlink ...
Or you could alternatively load the slot into the softlink... either 
way...  :)))

i.e.

ln -s slot_00 data

This is really cool stuff...


-- 

RE: Using multiple Tapechangers

[John Koenig]

Which changer glue are you using for this, please.

Are you splitting the magazine? i.e. one config uses tapes 0-7 and 
the other config uses tapes 8-14, with, possibly a cleaning tape on 
15?

Curious minds really want to know... :)

thx


At 11:35 AM -0400 7/24/02, Bort, Paul wrote:
>The configuration you're describing is not a problem, I've been doing it for
>two years now. Each different configuration has its own changer setup, and
>they can be completely independent.
>


-- 

MRU tape changer software from Compaq

[John Koenig]

Ref. an article from far, far away (Date: 2000/04/20):

(article quoted below)

I am trying to use the changer option with the MRU software from 
Compaq. This software has an entry point via the `robot` command and 
has syntax such these examples:

> robot load slot 0 drive 0
> robot unload drive 0 slot 0

Does anyone have any more information about configuring Amanda for 
use with an MRU changer utility? Is it true that the "changerfile" 
and "changerdev" variables are unnecessary? (as shown below)

I have a notion that simply providing a symlink to the robot utility 
in /Amanda_Installation_Location/libexec is overly simplistic and 
will probably not work as the interface Amanda expects may not be 
consistent with the MRU command interface.

Tips...? Please send them to me via private e-mail. (if you prefer) 
or send to the list if appropriate...

I really could use some help... 
Config files, etc...
I will post a summary of the entire installation process, to help 
others find answers in the future...


Thanx!

JF



>From: Nick Manka ([EMAIL PROTECTED])
>Subject: Re: Any good websites devoted to Amanda and DLT robotic tape changers
>
>
>In article <[EMAIL PROTECTED]>,
>   Edward Jong <[EMAIL PROTECTED]> writes:
>
>>  I used the media changer program and it works fine, but I am looking at the
>>  amanda.conf and tape changer examples that is supplied with amanda and I
>>  trying to figure them out.   I looking at working implementation of amanda
>>  but it is only using a regular tape drive (single, not robotic)
>
>Here is the relevant section of our amanda.conf, pretty short:
>
>runtapes 9 # number of tapes to be used in a single run of amdump
>tpchanger "chg-mru"# the tape-changer glue script
>tapedev "/dev/changer/nrmt0h"  # the no-rewind tape device to be used
>#changerfile "/var/adm/amanda/conf/.../changer.conf"
>#changerdev "/dev/mc24"
>
>(the script we use, chg-mru, can pull the changer[dev|file] options
>via libexec/getconf but I hardcoded them during testing). stc-changer
>seems to have all it's configuration either hard-coded into the
>script itself and or pulled from the stctl.conf on your machine.
>
>Ideally, all you need is to define the "tpchanger" and "runtapes"
>values and then place the stc-changer script into the Amanda libexec
>directory.

Re: sendbackup when client is the tape server

[John Koenig]

At 8:27 AM -0400 6/17/02, Lee Fellows wrote:
>What do you think is the difference whether sendbackup uses the loopback
>interface on the local machine or the external interface on the local
>machine? 
>My guess is that it depends on how you identify the client in
>your local hosts file: i.e., if the client has interfaces lo0 on
>127.0.0.1 and eth0 on 10.1.101.101 and you call your machine darkstar
>and associate darkstar to the 10.1.101.101 IP address in your hosts
>file, then, assuming nsswitch.conf does not direct lookups to DNS first,
>sendbackup would use eth0 if darkstar is your hostname in your disklist.
>If you wanted to force sendbackup to use the loopback interface on the
>local client, in the above scenario, associate darkstar to 127.0.0.1 and
>doublecheck that inetd/xinetd is listening to the loopback address as
>well as the external address.


I certainly do not identify my tapeserver as localhost in the disklist file.

So this would infer that I may be hogging the local LAN with about 40 GBs of traffic, 
for only the tapeserver's disks...

thanks for the consideration of this topic !

Regards,
J


PS: sorry about the lack of carriage returns... I am telling my mail client (Eudora 
5.x) to not wrap, but it seems to insist on wrapping anyway... Could an MTA be 
modifying the body-text, or is this more likely to insert CR's if indeed it was 
modifying my data... ??

specifying gnutar, trouble with OSF1 file system

[John Koenig]

Using AMANDA version 2.4.2p2 on OSF/Tru64 v5.1

On the client side, amanda is sending an inappropriate parameter to 
the /sbin/dump


=
sendsize: running "/sbin/dump 0Esf 1048576 - /net/home1"
running /usr/local/amanda-2.4.2p2/libexec/killpgrp
dump:
dump: Cannot open file-system file home1_dmn#home1_fs
dump: Bad file system specification or bad file system.  The raw device must
dump: be entered when the file system's pathname has not been entered in the
dump: fstab file.  A bad file system is reported when the the magic number
dump: is not found in the super block.
=


I think i should be using gnutar... As this would likely work-around 
this issue...

At configure time, I have specified:

--with-gnutar=/usr/local/gnu/bin/tar

Only problem, I have not figured out how to force "the use" of gnutar 
instead of the distribution's dump program. The client in this case 
is still calling dump? And, btw, would this be the same way I get 
amanda to use the "vdump" program? vdump was found at compile-time.

Though, when I get the configuration to work, will I again encounter 
amanda passing a bogus string which is likely obtained from the 
/etc/fstab file? I am specifying a file system location in the 
disklist:

host /path-to/home1  no-record


###

I believe it is OK to use no-record for testing right?
I read about this somewhere... I think...

Would it be weird to run the no-record config and see a
file system get dumped to the holding disk and then written to tape?
On another client -- which is working somewhat -- I saw this behavior...
What does no-record mean exactly?



thanks!!

sendbackup when client is the tape server (and nfs too)

[John Koenig]

This has probably been discussed before, however I am not successfully getting good 
hits on my searches on this topic...

I am curious how the tape server is sending it's backup data to itself when it is also 
the amanda client. Is it using the loopback device? Or is it putting this data onto 
the network? From the looks of the sendbackup debug file it looks similar to a regular 
client... like the data is going onto the network. 

(Apologies for me not setting up some a way to evaluate this for myself with packet 
watching sw, but I was hoping that someone that knows the answer could just document 
it back onto the mailing list)

The reason to ask this question is related to a scenario where I am trying to make a 
decision about how to include a particular file system in the disk list. For example, 
the file system to be added to the disklist just happens to be NFS mounted on the tape 
server. So should I insert a disk list entry with the hostname of the tape server 
(exercising the NFS code)? Or should I install an amanda client on the remote host 
(where this file system is local) and then add the file system to the disk list by way 
of the hostname of the remote host? 

Are there particular advantages to doing this one way or the other?

Thx for taking the time folks...
J


PS: (hmm maybe this has no impact on the sendbackup question above, though I am still 
intrested in the answer to that)

RE: ...but I didn't use that tape!

[John Koenig]

[re-listed from amanda-hackers]


>Chris,
>
>This might be more appropriate for amanda-users, but I'll give it a go:
>
>I don't see how a code change will help your hardware, and there are enough
>people on the lists on low- or no-budget networks that I won't suggest a new
>tape drive (yet). (Aside: clean the drive if you haven't lately, this sounds
>like dirty heads.)
>
>After each backup, I would suggest using amverify to determine if the tape
>is any good, then removing and re-labeling the tape if amverify fails. You
>said this 'doesn't appeal', but I think its cleanest when you've destroyed a
>good(?) backup on that media with a bad one. You need to get rid of the
>indexes for the tape regardless of what you do with the physical media.
>
>If you write a script to amverify/amrmtape/amlabel, you could make it log
>the re-labels to look for a pattern, such as a specific brand of tape
>causing the problem.


Does anyone have such a script to share? that is a:

   "script to amverify/amrmtape/amlabel"



>
>I would also be VERY suspicious of the 'good' backups on a drive that is
>sometimes 'bad'. If the data's that important, then a reliable backup drive
>should be equally important.
>
>Best of luck,
>Paul
>
>> -Original Message-
>> From: Chris Jones [mailto:[EMAIL PROTECTED]]
>> Sent: Wednesday, June 05, 2002 2:23 PM
>> To: [EMAIL PROTECTED]
>> Subject: ...but I didn't use that tape!
>>
>>
>> I have a lousy tape drive.  When I try to dump (or flush),
>> sometimes the
>> backups will fail on the first file (after the label) on the
>> tape.  In
>> other words, there's no valid backup data on the tape.  But Amanda
>> considers the tape to be used, and I can't put more data on
>> it unless I
>> either wait for the tape cycle to come around again, re-label
>> the tape
>> in question, or try and hack on the datafiles.  For various reasons,
>> none of these options really appeals to me.
>>
>> Would it be desirable to special-case this situation?  If a
>> tape write
>> fails before the first file has been completely written,
>> don't consider
>> the tape to have been used.
>>
>> Chris
>>
>> --
>> - [EMAIL PROTECTED]
>> Chris Jones   Mad
>> scientist at large
>>www.netbsd.org www.postgresql.org www.schemers.org www.python.org
>>

Re: changerfile for chg-mtx

[John Koenig]

HTH


There are usually some notes in the headers of each changer-glue script, too.

This is from "TAPE.CHANGERS"


chg-mtx: (former hp-changer)

A mtx-based tape changer script.  `changerdev' must specify the tape
device controlled by the mtx program, and `tapedev' must point to
no-rewind tape device to be used.  More than likely, both `changerdev'
and `tapedev' will reference the same device file. 


 `changerfile' must
specify a prefix for status files maintained by this script.  


If `changerfile' is defined as `/usr/adm/amanda/csd/changer', this script
will maintain files named `changer-clean' and `changer-access' in the
log directory.  You may have to edit the script to specify which slot
contains a cleaning tape (cleanslot).

The `mtx' program must support commands such as `-s', `-l' and `-u'.
If the one you've got requires `status', `load' and `unload', you
should use chg-zd-mtx instead, see below.



chg-zd-mtx:

Based on chg-mtx, but modified in order to support Zubkoff/Dandelion
version of mtx.  Eric DOUTRELEAU <[EMAIL PROTECTED]>, who
contributed this script, reported that it works on a Solaris/sparc box
with a HP 1557A stacker.

In addition to the `changerfile'-clean and the `changerfile'-access
files, it will maintain a `changerfile'-slot that indicates the
currently loaded slot.

amcheck: tapedev is /dev/null

[John Koenig]

Ugh...

I have broken my previously working configuration...

I was well on my way to adding more clients to a chg-multi configuration and for some 
"Good Reason" I rebuilt my installation from source. Since then it has not operated as 
well as it did before...

Actually the good reason was that realized I had built the server for an rsh/rhosts 
environment... (with the --without-amandahosts option enabled). And all the clients on 
this network are using .amandahosts, so I rebuilt it... with only this option 
removed... or so I thought...  starting having problems with amcheck seeing my 
chg-multi config... 

Now you and I know I have mucked something up but I way past the have been pulling my 
hair out trying to figure out what I done did... ;) 

So, why would amcheck be saying: WARNING: tapedev is /dev/null?

I see the man page says the amcheck uses /dev/null as the default for the tapedev 
parameter/directive. However, my amanda.conf has tapedev removed as I am expecting the 
changer glue, "prefix/libexec/chg-multi" to root out the tapedev from the changer 
config file at "configdir/DailySet1/chg-multi.conf". After all, I had this thing 
backing up to on of the two specified tape devices (slot 1 or slot 2) that were 
specified in the "cgh-multi.conf" file. 

Now, I can invoke calls to the changer script and get the appropriate output:

  prefix/libexec/chg-multi -reset

  prefix/libexec/chg-multi -info

  prefix/libexec/chg-multi -slot next

that all works...

Yet the amcheck is apparently not seeing the changer config...
Can anyone suggest why amcheck is telling me the tapedev is /dev/null?

Have I turned on a debug option or something that is forcing amcheck and amdump to 
default to a test or degraded mode?


Here is what amdump says when I attmpt to run things after amcheck complains of the 
"WARNING: tapedev is /dev/null"



*** A TAPE ERROR OCCURRED: [new tape not found in rack].
Some dumps may have been left in the holding disk.
Run amflush to flush them to tape.
The next tape Amanda expects to use is: a new tape.

FAILURE AND STRANGE DUMP SUMMARY:
  anton.medi /etc lev 0 FAILED [can't switch to incremental dump]
  skippy.med /etc lev 0 FAILED [can't switch to incremental dump]
  skippy.med /usr/local/etc lev 0 FAILED [can't switch to incremental dump]
  java.media /usr/local/etc lev 0 FAILED [can't switch to incremental dump]
  www20.medi /etc lev 0 FAILED [can't switch to incremental dump]
  java.media /etc lev 0 FAILED [can't switch to incremental dump]
  www20.medi /usr/local/etc lev 0 FAILED [can't switch to incremental dump]
  java.media /usr/local lev 0 FAILED [can't switch to incremental dump]




#


what other pertinent info could I post?
I gotta get this figured out...

thanks!!!
J

Re: ADIC FastStor with Quantum DLT7000 and chg-zd-mtx

[John Koenig]

>Hi there,
> 
>Is anybody using this setup with luck?

people do use chg-zd-mtx successfully.

i am not familiar with this tape robot to definitively know if the changer script, 
chg-zd-mtx, is appropriate.

Note that the changer script in AMANDA is not always the interface to the tape robot. 
Usually there is another program (such as mtx) the drives the robot and chg-zd-mtx 
drives MTX, as an example... I am not sure what you need, though.

>Can anybody tell me how to configure the changer.conf file?
> 

for starters...
read the ./doc/TAPE.CHANGERS document.
read the header in the  $prefix/libexec/chg-zd-mtx file.

if you are having trouble debugging after things are set up, then you can expect to 
get help from someone here, for sure...

Re: amanda - my way ?

[John Koenig]

I recommend either:

 a) spend REAL BIG money on adequate tape library with at least 10 8mm tapes (AIT-2, 
for example).

or 

 b) spend much smaller amount of money on large RAID-5 array of many many gigabytes 
using 3ware IDE card and cheap IDE disks... then back up to disk and forget the tapes. 
 

:)

Now go ride!


##

PS: To the amanda-users list... I could have sent my response via private e-mail 
however I thought by posting this opinion I would get some other opinions... This is 
my opinion after working with amanda for a while...  however I am far from an expert.

amanda binaries, NFS mounts, and setuid root

[John Koenig]

Is there a general consensus regarding the use of an NFS mount to make the AMANDA 
binaries available to the clients?

And if the consensus is, in certain situations, OK... Then can someone with more 
experience with network security specify what the acceptable conditions might be for 
allowing the NFS mount to be configured with the setuid root attribute...

And sorry to everyone who nows sees this is an NFS question. :)))

I suppose I could copy the amanda install tree to each architecture, Tru64 (RIP), 
Linux, Solaris, Irix, etc. But that is more hassle when it comes time to upgrade. But 
hey... by that time I'll be on my next job. :)))

thanks folks...
:)





 

Re: current-slot undefined in backward capable changer ?

[John Koenig]

Nevermind...

I found my answer...


>4.  SLOT NUMBERS AND THE "CURRENT" SLOT
>
>Some tape changers, such as carousels and gravity stackers, have a hardware
>notion of current position.  Others have no current position when no tape
>is loaded: all tapes are in their slots and the changer arm is docked away
>from the slots.
>
>Nevertheless, Amanda requires its tape-changer scripts to maintain the
>notion of a "current" position.  This is for performance reasons: as
>tapes tend to be loaded into the rack in order, and Amanda uses them
>in order, the next tape to use can be found much quicker if the
>position of the previous one is remembered.  As an example, the
>chg-multi.sh script maintains the current position in a
>"chg-multi.state" file (or any other file specified in a `statefile'
>line in the changer configuration file).

current-slot undefined in backward capable changer ?

[John Koenig]

OK... the changer spec, "TAPE.CHANGERS", sez:


##

 -info

 Outputs to stdout three fields: the current slot string, the
 number of slots, and whether the changer can go backwards (0
 if it can't, 1 if it can).  Same error handling as above.

Example:
   % chg-multi -info
   0 10 1# exit code returned is 0

   | |  |
   | |  |
   | |  +--- Backwards capable (boolean-like value)
   | |
   | +-- Number of slots for this configuration (int-like)
   |
   + Current Slot (int-like)

##

 From the John Jackson notes that exist on the site previously known 
as backupcentral, if the changer's configuration file specifies the 
gravity flag as false, then the changer can perform what is called 
"backwards" motion.

My question:

  1) Is the notion of a current-slot undefined for all changers that 
are capable of bidirectional (i.e. random access?) transport behavior 
(i.e. the changer can go backwards)?


  2) If current-slot is undefined, would it be OK to substitute "0" 
for the value of $curslot in the "-info" command? If not, what value 
should the changer script write out for the current_slot value when 
backwards is true?


Thanks again...
J

Elaboration on tape changer behavior

[John Koenig]

Can those with experience with changers elucidate some details for me?

What is the AMANDA/amtape policy when "rolling over" from the last 
tape to the first tape in my "Backup Set"? My config specifies the 
following:

dumpcycle 2 weeks
runspercycle 5
tapecycle 10 tapes

In my changer the slot numbering is zero-based. So I  have tapes in 
slots 0 thru 9.

Since I do not have a working changer script between amtape and my 
tape library manager sw layer, I am manually running robot load 
commands (etc...). I have three general questions:


  1) Will amtape manage the rollover from tape slot 9 to slot 0, and, 
in general, do the right thing? Is there any manual intervention on 
tape slot rollover? For example. must the admin tell AMANDA that it 
is OK to write over a tape (tape "00")?


  2) Is there any important information omitted from the document, 
"AMANDA Tape Changer Support"? Can it generally be trusted to specify 
the correct interfaces? I have found that some of the tape changer 
implementations in the AMANDA distribution do not correctly follow 
the API specifications in that document.


  3) Do people with single tape robots use multiple tape magazines; 
swapping them out after one tapecycle... as in Magazine A then 
Magazine B, then back to A? Or do some people just keep running tapes 
through the magazine, ad infinitum? Or do most people leave the same 
magazine installed until the tapes wear out? I presume this last 
question (3) is largely dependent upon one's goals... So if anybody 
cares to elaborate on changer methodologies used to achieve 
particular goals, then I would be grateful. Other will surely benefit 
as this is not an area that is discussed in much depth.

Thanks for your time...

J

Re: columnspec tweaks

[John Koenig]

Thanks Jon (LaBadie)...

RTFM again... AMANDA is more amazing the more I learn about it...

Thanks!

#


columnspec «string"
Defines the width of columns amreport should use. String is a comma 
(',') separated list of triples. Each triple consists of three parts 
which are separated by a equal sign ('=') and a colon (':') (see the 
example). These three parts specify:

+ the name of the column, which may be:

Compress (compression ratio)
Disk (client disk name)
DumpRate (dump rate in KBytes/sec) DumpTime (total dump time in 
hours:minutes) HostName (client host name)
Level (dump level)
OrigKB (original image size in KBytes) OutKB (output image size in 
KBytes) TapeRate (tape writing rate in KBytes/sec) TapeTime (total 
tape time in hours:minutes)

+ the amount of space
to display before the column (used to get whitespace between columns).
+ the width of the column itself.
If set to a negative value, the width will be calculated on demand to 
fit the largest entry in this column.

columnspec tweaks

[John Koenig]

I think the "columnspec" flexibility is a wonderful feature...

However... when the taper metrics for data throughput start reaching 
4-5 digits on the left side of the decimal point, the time strings in 
the same category (TAPER STATS) run into most significant digits of 
the throughput data...

Here is an example:


DUMP SUMMARY:
DUMPER STATS 
TAPER STATS
HOSTNAME   DISK L ORIG-KB   OUT-KB COMP% MMM:SS 
KB/s MMM:SS  KB/s
- 
--- 
redbelly   /net/redbelly/home1  0  617320   258752  41.9 
1:302883.3   0:495294.9
redbelly   /net/redbelly/lanl1  0  30   64 213.3   0:00 
0.0   0:005461.2
redbelly   /net/redbelly/lbnl1  16650  736  11.1   0:53 
12.7   0:0018841.4
ringneck   /net/ringneck/camb1  0  20   64 320.0   0:00 
0.0   0:009362.2
ringneck   /net/ringneck/tamu1  0  30   64 213.3   0:00 
59.1   0:09   7.2


|< widen window to here 
>|


I intend on looking at the code for columnspec, however I am in the 
process of implementing a changer script for my environment... so 
formatting will have to take a lower priority...

Primarily I wanted to document this behavior on the list...

thx...

General Changer Questions (chg-mtx)

[John Koenig]

In the files, chg-manual and chg-mtx there is a function in the code 
labelled loadslot(). In this funciton the dd command is used to look 
at the tape.

I have excerpted some code from chg-mtx, below...

This appears to be post-processing code following that event wherein a
tape is newly inserted into the drive...

Here are my questions:
--

1)  Is this entire section concerned with determining if the next 
tape is the cleaning tape? IOW, is the purpose of the DD to see if we 
are reading a cleaning tape?

3)  Also, when next-tape equals first-tape, what does AMANDA *want to 
do* when asked to write over the first tape in the magazine?

4)  I downloaded, built, and installed mtx-1.2.16rel, the SCSI 
control program available at SourceForge. I could not find certain 
options that are used in the chg-mtx below. For example, these 
options I do not find: -l, -u. Does anyone know the actions implied 
by:

mtx -u
mtx -l



Thank for any and all advice... I appreciate your time...

J


#
## from 3b3/libexec/chg-mtx
#

 # Slot 6 might contain an ordinary tape rather than a cleaning
 # tape. A cleaning tape auto-ejects; an ordinary tape does not.
 # We therefore have to read the status again to check what
 # actually happened.
 readstatus


 if [ $used -gt 0 ];then
   echo " -> unload $used" >> $logfile
   res=`$MTX -u $used`
   status=$?
   echo " -> status $status" >> $logfile
   echo " -> res$res" >> $logfile
   if [ $status -ne 0 ];then
 answer=" $myname: $res"
 code=2
 echo "Exit -> $answer" >> $logfile
 echo "$answer"
 exit $code
   fi
 fi
 if [ $whichslot = advance ];then
   answer="$load /dev/null"
   code=0
   echo "Exit -> $answer" >> $logfile
   echo "$answer"
   exit $code
 fi
 echo " -> load   $load" >> $logfile
 res=`$MTX -l $load`
 status=$?
 echo " -> status $status" >> $logfile
 echo " -> res$res" >> $logfile
 if [ $status -eq 0 ];then
   echo " -> rew $load" >> $logfile
   $MT $MTF $tape rewind
   $DD if=$tape bs=32k count=1 >> $logfile 2>&1
   answer="$load $tape"
   code=0
 else
   answer="$load $res"
   code=2
 fi
 echo "Exit -> $answer" >> $logfile
 echo "$answer"
 exit $code

###

WANTED: DDS-3 tape drive

[John Koenig]

Anyone have a DDS-3 (4mm, 12 GB native) unit that has been removed 
from service? I am looking for a unit that is considered obsolete by 
the current owner; which I hope would make it nearly worthless and 
therefore quite inexpensive to procure from "you"...  :)

Thank you...

"--disable-libtool" option in the 3b3 build

[John Koenig]

On a current version of Linux, building 3b3 with the configure option 
"--disable-libtool" works OK... It also builds fine without disabling 
libtool...

Can someone please suggest what I should examine to find the problem 
that is preventing an amanda 3b3 build on Tru64?

Whenever the "--disable-libtool" option is specified then it would 
appear to be used anyway. I have already determined that I am having 
problems building 3b3 on Tru64 when "--disable-libtool" is not 
specified... So I thought disabling the use of libtool would prevent 
any attempt to use libtool.

Any suggestions?

thx

libtool problem with amanda-2.4.3b3 build

[John Koenig]

I recently posted some build data about versions 2p2 and 3b3 with 
regard to the results of various configure options related to libtool 
as well as changing the version of libtool...

The 2p2 builds performed without a problem
when "--disable-libtool" was set or unset.

The 3b3 build dies with libtool errors REGARDLESS
of whether "--disable-libtool" is set or not.

Would this not indicate a problem with the build? Regardless of 
whether or not there is a problem with libtool on the Alpha/Tru64 
box?...

If one says "--disable-libtool" what should one expect?
A broken build with libtoool errors as the cuplrit?

Maybe my problem is not libtool when the "--disable-libtool" option 
is specified...

thx
dk

amanda-2.4.3b3 and libtool on Tru64 5.0

[John Koenig]

My issue is with libtool related failures when building 2.4.3b3

I tried several things suggested to me:

  o Am I running configure on a clean source tree?
YES, freshly unpacked.

  o Have I tried:

+ downgrading libtool?
YES to 1.3.4. Then re-intalled the 1.4.2 version again.
Results of build are consistently the same on both
  2.4.2p2 and 2.4.3b3 source trees, with either version.


+ upgrading libtool to beta version?
NO... All I could find was an alpha/development version.


+ specifying the --disable-libtool?
No need with 2.4.2p2 since this problem does not occur.
With 2.4.3b3, this allows compilation to not fail until
  later with a libtool error



Here is the output of 2.4.3b3 build with the --disable-libtool 
specified at the configure stage:

#

config.status: creating config/Makefile
config.status: creating Makefile
config.status: creating config/config.h
Making all in config
make[1]: Entering directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
cd .. \
   && CONFIG_FILES= CONFIG_HEADERS=config/config.h \
  /usr/bin/sh ./config.status
config.status: creating config/config.h
config.status: config/config.h is unchanged
make  all-am
make[2]: Entering directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
make[2]: Leaving directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
make[1]: Leaving directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
Making all in common-src
make[1]: Entering directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/common-src'
source='alloc.c' object='alloc.o' libtool=no \
depfile='.deps/alloc.Po' tmpdepfile='.deps/alloc.TPo' \
depmode=gcc3 /usr/bin/sh ../config/depcomp \
gcc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src 
-I/usr/local/gnu/include   -O2   -c `test -f alloc.c || echo 
'./'`alloc.c
source='amflock.c' object='amflock.o' libtool=no \
depfile='.deps/amflock.Po' tmpdepfile='.deps/amflock.TPo' \
depmode=gcc3 /usr/bin/sh ../config/depcomp \
gcc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src 
-I/usr/local/gnu/include   -O2   -c `test -f amflock.c || echo 
'./'`amflock.c
source='debug.c' object='debug.o' libtool=no \
depfile='.deps/debug.Po' tmpdepfile='.deps/debug.TPo' \
depmode=gcc3 /usr/bin/sh ../config/depcomp \
gcc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src 
-I/usr/local/gnu/include   -O2   -c `test -f debug.c || echo 
'./'`debug.c
source='dgram.c' object='dgram.o' libtool=no \
depfile='.deps/dgram.Po' tmpdepfile='.deps/dgram.TPo' \
depmode=gcc3 /usr/bin/sh ../config/depcomp \
gcc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src 
-I/usr/local/gnu/include   -O2   -c `test -f dgram.c || echo 
'./'`dgram.c
dgram.c: In function `dgram_recv':
dgram.c:292: warning: passing arg 6 of `recvfrom' from incompatible 
pointer type
source='error.c' object='error.o' libtool=no \
depfile='.deps/error.Po' tmpdepfile='.deps/error.TPo' \
depmode=gcc3 /usr/bin/sh ../config/depcomp \
gcc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src 
-I/usr/local/gnu/include   -O2   -c `test -f error.c || echo 
'./'`error.c
source='file.c' object='file.o' libtool=no \
depfile='.deps/file.Po' tmpdepfile='.deps/file.TPo' \
depmode=gcc3 /usr/bin/sh ../config/depcomp \
gcc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src 
-I/usr/local/gnu/include   -O2   -c `test -f file.c || echo 
'./'`file.c
source='fileheader.c' object='fileheader.o' libtool=no \
depfile='.deps/fileheader.Po' tmpdepfile='.deps/fileheader.TPo' \
depmode=gcc3 /usr/bin/sh ../config/depcomp \
gcc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src 
-I/usr/local/gnu/include   -O2   -c `test -f fileheader.c || echo 
'./'`fileheader.c
sh ./../regex-src/mkh -o -i _REGEX_H_ ./../regex-src/regex2.h 
./../regex-src/regcomp.c ./../regex-src/regexec.c 
./../regex-src/regerror.c ./../regex-src/regfree.c >regex.h
source='match.c' object='match.o' libtool=no \
depfile='.deps/match.Po' tmpdepfile='.deps/match.TPo' \
depmode=gcc3 /usr/bin/sh ../config/depcomp \
gcc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src 
-I/usr/local/gnu/include   -O2   -c `test -f match.c || echo 
'./'`match.c
source='protocol.c' object='protocol.o' libtool=no \
depfile='.deps/protocol.Po' tmpdepfile='.deps/protocol.TPo' \
depmode=gcc3 /usr/bin/sh ../config/depcomp \
gcc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src 
-I/usr/local/gnu/include   -O2   -c `test -f protocol.c || echo 
'./'`protocol.c
sh ./../regex-src/mkh -o -p ./../regex-src/regcomp.c >regcomp.ih
source='regcomp.c' object='regcomp.o' libtool=no \
depfile='.deps/regcomp.Po' tmpdepfile='.deps/regcomp.TPo' \
depmode=gcc3 /usr/bin/sh ../config/depcomp \
gcc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src 
-I/usr/local/gnu/include   -O2   -c `test -f regcomp.c || echo 
'./'`regcomp.c
sh ./../regex-src/mkh -o -p ./../regex-src/regerror.c >regerror.ih
source='regerror.c' object='regerror.o' libtool=no \
d

Re: amanda 2.4.3b3

[John Koenig]

Hello again...

I took the advice and inserted "--disable-libtool" into my configure.
(OSF1 V5.1 732 alpha)

Files actually started to compile this time instead of dying right at the 
beginning.. 
This time... not long after some files *did* compile, it seems the broken libtool 
invocation showed up again... How can that be? I said 'disable-libtool'.

Does this information plus the fact that the 2.4.2p2 build worked fine without the 
disabling the use of libtool, provide any useful information to the maintainers?

###

Here is my output from the most recent build attempt of amanda-2.4.3b3:

config.status: creating Makefile
config.status: creating config/config.h
Making all in config
make[1]: Entering directory `/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
cd .. \
  && CONFIG_FILES= CONFIG_HEADERS=config/config.h \
 /usr/bin/sh ./config.status
config.status: creating config/config.h
config.status: config/config.h is unchanged
make  all-am
make[2]: Entering directory `/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
make[2]: Leaving directory `/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
make[1]: Leaving directory `/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
Making all in common-src
make[1]: Entering directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/common-src'
source='alloc.c' object='alloc.o' libtool=no \
depfile='.deps/alloc.Po' tmpdepfile='.deps/alloc.TPo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
cc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src  -I/usr/local/gnu/include   
-O2   -c `test -f alloc.c || echo './'`alloc.c
source='amflock.c' object='amflock.o' libtool=no \
depfile='.deps/amflock.Po' tmpdepfile='.deps/amflock.TPo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
cc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src  -I/usr/local/gnu/include   
-O2   -c `test -f amflock.c || echo './'`amflock.c
source='debug.c' object='debug.o' libtool=no \
depfile='.deps/debug.Po' tmpdepfile='.deps/debug.TPo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
cc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src  -I/usr/local/gnu/include   
-O2   -c `test -f debug.c || echo './'`debug.c
source='dgram.c' object='dgram.o' libtool=no \
depfile='.deps/dgram.Po' tmpdepfile='.deps/dgram.TPo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
cc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src  -I/usr/local/gnu/include   
-O2   -c `test -f dgram.c || echo './'`dgram.c
cc: Warning: dgram.c, line 110: In this statement, the referenced type of the pointer 
value "&len" is "unsigned int", which is not compatible with "int" because they differ 
by signed/unsigned attribute. (ptrmismatch1)
if(getsockname(s, (struct sockaddr *)&name, &len) == -1) {
^
cc: Warning: dgram.c, line 292: In this statement, the referenced type of the pointer 
value "&addrlen" is "unsigned long", which is not compatible with "int". (ptrmismatch)
(struct sockaddr *)fromaddr, &addrlen);
-^
source='error.c' object='error.o' libtool=no \
depfile='.deps/error.Po' tmpdepfile='.deps/error.TPo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
cc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src  -I/usr/local/gnu/include   
-O2   -c `test -f error.c || echo './'`error.c
source='file.c' object='file.o' libtool=no \
depfile='.deps/file.Po' tmpdepfile='.deps/file.TPo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
cc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src  -I/usr/local/gnu/include   
-O2   -c `test -f file.c || echo './'`file.c
source='fileheader.c' object='fileheader.o' libtool=no \
depfile='.deps/fileheader.Po' tmpdepfile='.deps/fileheader.TPo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
cc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src  -I/usr/local/gnu/include   
-O2   -c `test -f fileheader.c || echo './'`fileheader.c
source='match.c' object='match.o' libtool=no \
depfile='.deps/match.Po' tmpdepfile='.deps/match.TPo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
cc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src  -I/usr/local/gnu/include   
-O2   -c `test -f match.c || echo './'`match.c
source='protocol.c' object='protocol.o' libtool=no \
depfile='.deps/protocol.Po' tmpdepfile='.deps/protocol.TPo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
cc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src  -I/usr/local/gnu/include   
-O2   -c `test -f protocol.c || echo './'`protocol.c
source='regcomp.c' object='regcomp.o' libtool=no \
depfile='.deps/regcomp.Po' tmpdepfile='.deps/regcomp.TPo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
cc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src  -I/usr/local/gnu/include   
-O2   -c `test -f regcomp.c || echo './'`regcomp.c
source='regerror.c' object='regerror.o' libtool=no \
depfile='.deps/regerror.Po' tmpdepfile='.deps/regerror.TPo' \
depmode=tru64 /usr/bin/sh ../config/depc

Re: amanda 2.4.3b3

[John Koenig]

Thank you for all your (and everyone else involved) hard work...

Though... I am having build issues...

Can someone suggest soemething that I should examine (and fix) to 
make the build work properly?  thank you...  -JFK

#

config.status: creating Makefile
config.status: creating config/config.h
Making all in config
make[1]: Entering directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
cd .. \
   && CONFIG_FILES= CONFIG_HEADERS=config/config.h \
  /usr/bin/sh ./config.status
config.status: creating config/config.h
config.status: config/config.h is unchanged
make  all-am
make[2]: Entering directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
make[2]: Leaving directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
make[1]: Leaving directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/config'
Making all in common-src
make[1]: Entering directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/common-src'
source='alloc.c' object='alloc.lo' libtool=yes \
depfile='.deps/alloc.Plo' tmpdepfile='.deps/alloc.TPlo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
/usr/bin/sh ../libtool --mode=compile cc -DHAVE_CONFIG_H -I. -I. 
-I../config -I./../regex-src  -I/usr/local/gnu/include   -O2   -c -o 
alloc.lo `test -f alloc.c || echo './'`alloc.c
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
make[1]: *** [alloc.lo] Error 1
make[1]: Leaving directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b3/common-src'
make: *** [all-recursive] Error 1


#



>The Amanda core team is pleased to announce the release of Amanda
>2.4.3b3. It is the third beta release of 2.4.3.  All known bugs
>are fixed.
>

amanda-2.4.3b2-20020306 build failed

[John Koenig]

Hello,

On a

OSF1 V5.1 732 alpha

I am getting the following output from an attempted build.
Can someone suggest something for me to examine to make this error go away?
I have made sure that I have a recent GNU libtool in my path.

thanks
-JK




.
.
.
config.status: creating Makefile
config.status: creating config/config.h
Making all in config
make[1]: Entering directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b2-20020306/config'
cd .. \
   && CONFIG_FILES= CONFIG_HEADERS=config/config.h \
  /usr/bin/sh ./config.status
config.status: creating config/config.h
config.status: config/config.h is unchanged
make  all-am
make[2]: Entering directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b2-20020306/config'
make[2]: Leaving directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b2-20020306/config'
make[1]: Leaving directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b2-20020306/config'
Making all in common-src
make[1]: Entering directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b2-20020306/common-src'
source='alloc.c' object='alloc.lo' libtool=yes \
depfile='.deps/alloc.Plo' tmpdepfile='.deps/alloc.TPlo' \
depmode=tru64 /usr/bin/sh ../config/depcomp \
/usr/bin/sh ../libtool --mode=compile cc -DHAVE_CONFIG_H -I. -I. 
-I../config -I./../regex-src  -I/usr/local/gnu/include   -O2   -c -o 
alloc.lo `test -f alloc.c || echo './'`alloc.c
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
../libtool: print: not found
make[1]: *** [alloc.lo] Error 1
make[1]: Leaving directory 
`/net/redbelly/scratch1/Admin/Build/amanda-2.4.3b2-20020306/common-src'
make: *** [all-recursive] Error 1


#

Using cvs to checkout amanda

[John Koenig]

Here are the instructions from the sourceforge page:



###

cvs -d:pserver:[EMAIL PROTECTED]:/cvsroot/amanda login

cvs -z3 
-d:pserver:[EMAIL PROTECTED]:/cvsroot/amanda co 
modulename

###

My question is: what module name do I use? "amanda-2" ?

I have the following choices:

  NSRcrypt/
  amanda/
  amanda-2/
  amanda-krb-2/
  amcore/ 
  www/

client server nfs question

[John Koenig]

Is it correct that an AMANDA client (on the same host as the AMANDA 
tape server) will still use the mechanisms as a remote host 
(inetd.conf and amandad program) to communicate with the tape server? 
I believe this to be yes...

If the tape server also has lots of NFS mounts of some of the same 
file systems on remote hosts that require backup, would it be best to 
specify the tapeserver hostname in the disklist by using the file 
system specifiers that utilize NFS mounts? Or would it be better to 
specify the remote host in the disklist, pulling NFS out of the 
equation? Or is this 6 one way and half-dozen another way?

thx
J

Re: Installing amanda on TRU64 5.1

[John Koenig]

At 5:10 PM +1300 2/12/02, Clinton Dilks wrote:
>extern int ruserok P((const char *rhost, int suser,
>---^
>cc: Error: amanda.h, line 946: In this declaration, the type of
>"ruserok" is not compatible with the type of a previous declaration of
>"ruserok" at line number 705 in file /usr/include/unistd.h. (notcompat)
>extern int ruserok P((const char *rhost, int suser,
>---^


John made this in response to my encounter with this, recently...

The fix is probably already in the cvs tree (I speculate)

This has the fix to the configure script so you will not have to fart 
around commenting out the corresponding piece of the amanda.h file.

   

Re: specifying gnutar, trouble with OSF1 file system

[John Koenig]

A lot of my issues with the client and dump were resolved by using 
the "advfs.diff" patch located here: 


Question: Have all these patches been rolled into post-2.4.2p2 builds?

"All the sudden" amanda and vdump and everything is just working like 
a charm on OSF1 (Tru64) v5.1

Though, I am still interested in forcing the use of tar (gnutar 1.13).

Currently, I have this in my configuration however it is calling vdump:

--with-gnutar=/usr/local/gnu/bin/tar

This is after the advfs.diff patch was applied.

Question: How do I force the dump program to use gnutar?

Oh... and the no-record option does perform the backup however with 
this dumptype specified the "backup command" (e.g. dump, vdump, etc) 
will not update its state file (e.g. /etc/dumpdates)



>Using AMANDA version 2.4.2p2 on OSF/Tru64 v5.1
>
>On the client side, amanda is sending an inappropriate parameter to 
>the /sbin/dump
>
>
>=
>sendsize: running "/sbin/dump 0Esf 1048576 - /net/home1"
>running /usr/local/amanda-2.4.2p2/libexec/killpgrp
>dump:
>dump: Cannot open file-system file home1_dmn#home1_fs
>dump: Bad file system specification or bad file system.  The raw device must
>dump: be entered when the file system's pathname has not been entered in the
>dump: fstab file.  A bad file system is reported when the the magic number
>dump: is not found in the super block.
>=
>
>
>I think i should be using gnutar... As this would likely work-around 
>this issue...
>
>At configure time, I have specified:
>
>   --with-gnutar=/usr/local/gnu/bin/tar
>
>Only problem, I have not figured out how to force "the use" of 
>gnutar instead of the distribution's dump program. The client in 
>this case is still calling dump? And, btw, would this be the same 
>way I get amanda to use the "vdump" program? vdump was found at 
>compile-time.
>
>Though, when I get the configuration to work, will I again encounter 
>amanda passing a bogus string which is likely obtained from the 
>/etc/fstab file? I am specifying a file system location in the 
>disklist:
>
>   host /path-to/home1  no-record
>
>
>###
>
>I believe it is OK to use no-record for testing right?
>I read about this somewhere... I think...
>
>Would it be weird to run the no-record config and see a
>file system get dumped to the holding disk and then written to tape?
>On another client -- which is working somewhat -- I saw this behavior...
>What does no-record mean exactly?
>
>
>
>thanks!!

specifying gnutar, trouble with OSF1 file system

[John Koenig]

Using AMANDA version 2.4.2p2 on OSF/Tru64 v5.1

On the client side, amanda is sending an inappropriate parameter to 
the /sbin/dump


=
sendsize: running "/sbin/dump 0Esf 1048576 - /net/home1"
running /usr/local/amanda-2.4.2p2/libexec/killpgrp
dump:
dump: Cannot open file-system file home1_dmn#home1_fs
dump: Bad file system specification or bad file system.  The raw device must
dump: be entered when the file system's pathname has not been entered in the
dump: fstab file.  A bad file system is reported when the the magic number
dump: is not found in the super block.
=


I think i should be using gnutar... As this would likely work-around 
this issue...

At configure time, I have specified:

--with-gnutar=/usr/local/gnu/bin/tar

Only problem, I have not figured out how to force "the use" of gnutar 
instead of the distribution's dump program. The client in this case 
is still calling dump? And, btw, would this be the same way I get 
amanda to use the "vdump" program? vdump was found at compile-time.

Though, when I get the configuration to work, will I again encounter 
amanda passing a bogus string which is likely obtained from the 
/etc/fstab file? I am specifying a file system location in the 
disklist:

host /path-to/home1  no-record


###

I believe it is OK to use no-record for testing right?
I read about this somewhere... I think...

Would it be weird to run the no-record config and see a
file system get dumped to the holding disk and then written to tape?
On another client -- which is working somewhat -- I saw this behavior...
What does no-record mean exactly?



thanks!!

Re: config/build assumptions (and how to fix)

[John Koenig]

>OK, time for "Configure script debugging 101" :-).


This is awesome, John!!
Thanks for taking the time... You rock!



>
>>Config.log shows the following:
>>
>>configure:4003: checking for perl5
>  >configure:4003: checking for perl
>

[snip]

>If you're just going to concentrate on the perl code, add "exit 0"
>as well.  That will stop the script at this point and speed up debugging.


So, just focusing on locating perl version 5, for the moment...


>Now, remove config.cache and rerun ./configure, then look at the log (or
>what comes out on the terminal) and figure out what's going wrong:
>
>   * Did LOCSYSPATH have everything you expected it to, in particular,
> /usr/bin?


Yes.

>
>   * Did the shell loop properly though all the PATH elements, again, in
> particular, /usr/bin?


It never found a file labelled "perl5"

But, *Yes*, the loop terminated when a match was found on 'perl'.
It matched on a path that came before /usr/bin in the PATH variable.
That path is /usr/local/bin. An instance of 'perl' (sans the 
character '5') was found. However, if one runs this instance of perl, 
one sees that it is
indeed version 5  (v5.005_03).


>
>   * Did the "ls -lL" show a match that should have happened?


Yep!


>
>   * Did the script say "Found" but yet it continued to loop?


Yes:
Found "/usr/local/bin/perl"


So, it would appear that the script determines version numbers based 
on file names...


I'll go on to use these same strategies to see what is happening with 
more important facilities, like file locking, which configure thinks 
is non-existent.

dk

Re: config/build assumptions (and how to fix)

[John Koenig]

Oopsie!

I forgot the most important error of the bunch

the file locking...

Why is configure missing these on OSF1 v5.1?

 From config.log

configure:9534: checking for fflush declaration in stdio.h
configure:9593: checking for flock
configure:9621: gcc -o conftest  -I/usr/local/gnu/include 
-I/usr/local/gnu/include/readline   conftest.c -lm -ltermcap\
  1>&5
configure:9647: checking for flock declaration in sys/file.h
configure:9706: checking for fprintf


configure:19224: checking whether posix fcntl locking works
configure:19244: gcc -o conftest  -I/usr/local/gnu/include 
-I/usr/local/gnu/include/readline   conftest.c -lm -ltermcap 
\
1>&5

Re: config/build assumptions (and how to fix)

[John Koenig]

Here is some data from the config.log on OSF1 V5.1 732 alpha
Note that this is a newly purchased computer so that 5.1 distribution
on the disk's images is relatively pristine...  i.e. not a lot
cooks in this kitchen... just two.


>
>>checking for perl5... no
> >
> >   [Note perl 5.005 is at /usr/bin/perl...
>>   whereis finds it no problem...
> >   /usr/bin is in my path
>
>Don't know why this happened.  There should be more in config.log.



Config.log shows the following:

configure:4003: checking for perl5
configure:4003: checking for perl




>
>>checking for yywrap in -lfl... no
>>
>>
>>   [Note: I have installed the GNU flex-2.5.4
>>  thinking the yywrap would be included...
>>  The fl library file is where I expect.
>
>Ditto.



configure:5669: checking for flex
configure:5702: checking for flex
configure:5736: checking for yywrap in -lfl
configure:5755: gcc -o conftest  -I/usr/local/gnu/include 
-I/usr/local/gnu/include/readline   conftest.c -lfl  1>&5
/usr/bin/ld:
Can't locate file for: -lfl
collect2: ld returned 1 exit status
configure: failed program was:
#line 5744 "configure"
#include "confdefs.h"

[what follows and is not included here: about 1500 lines of C code]

Do you need to see this code, John?

I am suspecting that my problem is that I am not calling a gnu linker




>
>>checking for history.h... no
>>checking for readline.h... no
>>checking for readline/history.h... no
>>checking for readline/readline.h... no
>>
>>   [Note: These are installed and accessible.
>>  Why not found ??
>
>Ditto.



configure:6259: checking for readline.h
configure:6269: gcc -E conftest.c >/dev/null 2>conftest.out
configure:6265:22: readline.h: No such file or directory
configure: failed program was:
#line 6264 "configure"
#include "confdefs.h"
#include 
configure:6259: checking for readline/history.h
configure:6269: gcc -E conftest.c >/dev/null 2>conftest.out
configure:6265:30: readline/history.h: No such file or directory
configure: failed program was:
#line 6264 "configure"
#include "confdefs.h"
#include 
configure:6259: checking for readline/readline.h
configure:6269: gcc -E conftest.c >/dev/null 2>conftest.out
configure:6265:31: readline/readline.h: No such file or directory
configure: failed program was:
#line 6264 "configure"
#include "confdefs.h"
#include 




>
>If there is something basic wrong with your compiler (for instance, it
>requires a license and anything you do generates an error), ./configure
>is going to say lots of things were not found when what it really means
>is the test failed for some other reason.


We do have licensed compilers, true, but I am not using them...
My current invocation PATH and LD_LIBRARY_PATH variables are:


PATH=/usr/local_cci/gcc-3.0.3/bin:/usr/local/gnu/bin:.:/net/cci/dks/bin:/net/cci/dks/bin/ALPHA:/usr/local_cci/bin:/usr/local/bin:/usr/bin:/usr/sbin:/usr/dt/bin:/usr/bin/X11:/usr/bin/X11/demos:/etc:/usr/etc

LD_LIBRARY_PATH=/usr/local_cci/gcc-3.0.3/lib


osf1host> gcc -v
Reading specs from 
/usr/local_cci/gcc-3.0.3/lib/gcc-lib/alphaev67-dec-osf5.1/3.0.3/specs
Configured with: ../gcc-3.0.3/configure --prefix=/usr/local_cci/gcc-3.0.3 
--enable-languages=c,c++,f77
Thread model: single
gcc version 3.0.3

Note that the following directories are NFS mounts:

/usr/local/gnu
/usr/local_cci

Is it possible that AMANDA's build programs may not be dealing well with the way
the symlinks are resolved to their actual paths instead of the paths that are set
in the PATH variable...

If you `cd /usr/local/gnu` and then do a `pwd` you are actually in
this directory:

/tmp_mnt/net/boa/usr_local/gnu

I noticed Linux systems preserve these faux pathnames consistently
but other platforms, typically the non-Free ones, do not.

Compared to a Linux system:

LinuxHOST> cd /usr/local_cci
LinuxHOST> pwd
/usr/local_cci


Is this a problem...?


thanks ALL!

Re: Another amcheck problem

[John Koenig]

Glad you got your stuff working Shawn...

I was going to suggest comparing the IP address of your tape server 
with that of a reverse DNS lookup  I think I had problems with 
that

So while I am thinking about it, I will pose a question
on this same thread since the the diagnostic errors
seem similar

I configured a client with an IP address in the ".amandahosts" file.
The ip address was a "static" ip address that is associated with a
gateway. The tape server is on a private (192.168.1) network with an
IP address of 192.168.1.11.

With only the IP address of the gateway, I could not get through the 
BSD style authentication...

Then I changed the contents of the .amandahosts" file to the
actual reverse DNS name served by DNS servers and things started
working... (which is associate to the same IP address)

Now, does this seems reasonable or did I also probably change
some other thing in the process that corrected my auth problem,
unbeknownst to me... :)

J

config/build assumptions (and how to fix)

[John Koenig]

Is it a bad assumption to think that amanda's build process will 
"bail" if there is something that it needs which is not found by the 
configure/autoconf programs?


Here are some samples of the more prominent complaints:

[ from ALPHA running OSF1 v5.1 ]

creating cache ./config.cache
checking host system type... alphaev6-dec-osf5.1
checking target system type... alphaev6-dec-osf5.1
checking build system type... alphaev6-dec-osf5.1
.
.
checking for working aclocal... missing
checking for working autoconf... missing
checking for working automake... missing
checking for working autoheader... missing
checking for working makeinfo... missing
.
checking for bison... no
checking for byacc... no
.
checking for perl5... no

   [Note perl 5.005 is at /usr/bin/perl...
   whereis finds it no problem...
   /usr/bin is in my path

checking for yywrap in -lfl... no


   [Note: I have installed the GNU flex-2.5.4
  thinking the yywrap would be included...
  The fl library file is where I expect.


checking for history.h... no
checking for readline.h... no
checking for readline/history.h... no
checking for readline/readline.h... no

   [Note: These are installed and accessible.
  Why not found ??

.
.
checking whether posix fcntl locking works... no
checking whether flock locking works... no
checking whether lockf locking works... no
checking whether lnlock locking works... no
configure: warning: *** No working file locking capability found!
configure: warning: *** Be VERY VERY careful.


#

What can safely be ignored...

aclocal?
autoconf?
automake?
makeinfo?
bison?
byacc?
yywrap?


I understand the file locking is required. I just got to get the 
configure script to see it...

###

My attempts to tell the configure program where the includes and 
libraries are, look like this:

>  cd amanda-2.4.2p2

>  CFLAGS="-I/usr/local/include -I/usr/local/include/readline"   \
LD_LIBRARY_PATH="/usr/local/lib:$LD_LIBRARY_PATH"   \
./configure \
 --prefix=/usr/local/amanda-2.4.2p2  \
 --without-amandahosts   \
 --with-user=amanda  \
 --with-group=operator   \
 --with-configdir=/usr/local/etc/amanda  \
 --with-tmpdir=/tmp/amanda   \
 --with-debugging\
 --with-debug-days=4


###

Thanks everyone

-J

Re: "No working file locking" and ruserok compile error

[John Koenig]

>
>Take a look at config.log and see why the test programs are failing.
>It might even be the ruserok() problem.
>
>>B) When I compile on OSF/1 v5.1 (or DEC ALPHA, or Tru64), I get an
>> error about ruserok being declared twice...
>> I am using a gnu make 3.79 and gcc 3.0.3.
>> Can I comment out the ruserok declaration in amanda.h with no
>>  adverse effects?
>
>In theory, the declaration in amanda.h is only enabled if ./configure
>did **not** find a declaration in a certain set of system header files.
>Where is the second (system) declaration on your system?  This may
>just be a matter of adding another header file to the list or fixing a
>test program.  Again, look for what happened in config.log.
>


I have searched the config.log file and I see no references to the 
string 'ruserok' with regard to the "multiply declared ruserok" 
issue What other information could I provide about the ruserok 
issue? What can I do to debug it, other than the config.log file?

and...

I see no reference to the flock tests or any lock tests for that 
matter... The only occurrence of the string 'lock' in the config.log 
file was inside the word 'block'

Where else can I look to see why the configure script is not finding 
evidence of file locking on a OSF/1 v5.1 Unix system?

##

Since this is a proprietary vendor supplied distribution -- unlike 
the GNU distribution -- readline was also missing. So, since I 
thought this to be an easy fix, I built and installed readline, 
however subsequent runs of the AMANDA configure script did not pick 
it up... Maybe someone can make a suggestion to make it work? Here is 
what I am doing to set the environment in which configure runs:


CFLAGS="-I/usr/local/include"   \
LD_LIBRARY_PATH="/usr/local/lib:$LD_LIBRARY_PATH"   \
./configure \
 --prefix=/usr/local/amanda-2.4.2p2  \
 --without-amandahosts   \
 --with-user=amanda  \
 --with-group=operator   \
 --with-configdir=/usr/local/etc/amanda  \
 --with-tmpdir=/tmp/amanda   \
 --with-debugging\
 --with-debug-days=1 \


previous to the above invocation:
CFLAGS was undefined
LD_LIBRARY_PATH=/usr/local/gcc-3.0.3/lib

The readline hdr files and lib were/are here:

 /usr/local/include/readline/*.h

 /usr/local/lib/libreadline.a

Thanks to all in this wonderful community
May all your backups be bit-perfect.

.amandahosts

[John Koenig]

Is there a way to determine what configuration options were compiled 
into a particular amanda build? Is this the header information in the 
top of an amandad.debug file?  If so, then where is there indication 
about '.amandahosts'?

I have re-configured with the WITHOUT-amanda-hosts  option and I 
am still getting amandad debug files indicating an ERROR due to:

"open of //.amandahosts failed"


which shows that, I do not have a home directory for amanda on this 
client... My environment is on where rsh and hosts.equiv are 
routinely used.

I did not think a home directory for the amanda user was necessary on 
the amanda clients...  especially if one was building for use without 
the amandahosts file...

I am very sure -- after about 3 re-compiles on both the server (tape 
host) and the client -- that the amandahosts option should not be 
enabled

What am I missing?

thanks...
jfk

"No working file locking" and ruserok compile error

[John Koenig]

Two issues I need some perspective on...


A) What should I do about warnings regarding an absence of file locking
Could this create run-time problems and be a hindrance to success?


checking disk device prefixes... /dev/ - /dev/r
checking whether posix fcntl locking works... no
checking whether flock locking works... no
checking whether lockf locking works... no
checking whether lnlock locking works... no
configure: warning: *** No working file locking capability found!
configure: warning: *** Be VERY VERY careful.





B) When I compile on OSF/1 v5.1 (or DEC ALPHA, or Tru64), I get an
error about ruserok being declared twice...
I am using a gnu make 3.79 and gcc 3.0.3.
Can I comment out the ruserok declaration in amanda.h with no
adverse effects?

Uhm: Note to maintainers... the .3beta builds on source forge
 fail worse on the Alphas compared to this trivial ruserok
 issue on the 2.4.2p2 build.


jfk@host> gcc -v
Reading specs from 
/usr/local_cci/gcc-3.0.3/lib/gcc-lib/alphaev67-dec-osf5.1/3.0.3/specs
Configured with: ../gcc-3.0.3/configure 
--prefix=/usr/local_cci/gcc-3.0.3 --enable-languages=c,c++,f77
Thread model: single
gcc version 3.0.3



jfk@host> ./configure   \
 --without-server\
 --prefix=/usr/local/amanda-2.4.2p2  \
 --with-user=amanda  \
 --with-group=operator   \
 --with-tmpdir=/tmp/amanda   \
 --with-debugging\
 --with-debug-days=1

.
.
.
.
creating server-src/Makefile
creating server-src/amstatus.pl
creating tape-src/Makefile
creating config/Makefile
creating Makefile
creating config/config.h



jfk@host> make

Making all in config
make[1]: Entering directory `/net/Admin/Build/amanda-2.4.2p2/config'
make[1]: Leaving directory `/net/Admin/Build/amanda-2.4.2p2/config'
Making all in common-src
make[1]: Entering directory `/net/Admin/Build/amanda-2.4.2p2/common-src'
/usr/bin/sh ../libtool --mode=compile gcc -DHAVE_CONFIG_H -I. -I. 
-I../config -I./../regex-src-g -O2   -c alloc.c
gcc -DHAVE_CONFIG_H -I. -I. -I../config -I./../regex-src -g -O2 -c 
alloc.c -o alloc.o
In file included from alloc.c:33:
amanda.h:947: conflicting types for `ruserok'
/usr/local/gcc-3.0.3/lib/gcc-lib/alphaev67-dec-osf5.1/3.0.3/include/unistd.h:714: 
previous declaration of `ruserok'
make[1]: *** [alloc.lo] Error 1
make[1]: Leaving directory `/net/Admin/Build/amanda-2.4.2p2/common-src'
make: *** [all-recursive] Error 1

MRU tape changer software from Compaq

[John Koenig]

Ref. an article from far, far away (Date: 2000/04/20):

(article quoted below)

I am trying to use the changer option with the MRU software from 
Compaq. This software has an entry point via the `robot` command and 
has syntax such these examples:

> robot load slot 0 drive 0
> robot unload drive 0 slot 0

Does anyone have any more information about configuring Amanda for 
use with an MRU changer utility? Is it true that the "changerfile" 
and "changerdev" variables are unnecessary? (as shown below)

I have a notion that simply providing a symlink to the robot utility 
in /Amanda_Installation_Location/libexec is overly simplistic and 
will probably not work as the interface Amanda expects may not be 
consistent with the MRU command interface.

Tips...? Please send them to me via private e-mail. (if you prefer) 
or send to the list if appropriate...

I really could use some help... 
Config files, etc...
I will post a summary of the entire installation process, to help 
others find answers in the future...


Thanx!

JF



>From: Nick Manka ([EMAIL PROTECTED])
>Subject: Re: Any good websites devoted to Amanda and DLT robotic tape changers
>
>
>In article <[EMAIL PROTECTED]>,
>   Edward Jong <[EMAIL PROTECTED]> writes:
>
>>  I used the media changer program and it works fine, but I am looking at the
>>  amanda.conf and tape changer examples that is supplied with amanda and I
>>  trying to figure them out.   I looking at working implementation of amanda
>>  but it is only using a regular tape drive (single, not robotic)
>
>Here is the relevant section of our amanda.conf, pretty short:
>
>runtapes 9 # number of tapes to be used in a single run of amdump
>tpchanger "chg-mru"# the tape-changer glue script
>tapedev "/dev/changer/nrmt0h"  # the no-rewind tape device to be used
>#changerfile "/var/adm/amanda/conf/.../changer.conf"
>#changerdev "/dev/mc24"
>
>(the script we use, chg-mru, can pull the changer[dev|file] options
>via libexec/getconf but I hardcoded them during testing). stc-changer
>seems to have all it's configuration either hard-coded into the
>script itself and or pulled from the stctl.conf on your machine.
>
>Ideally, all you need is to define the "tpchanger" and "runtapes"
>values and then place the stc-changer script into the Amanda libexec
>directory.