Re: Transport encryption

[Kamil Jońca]

Marc Schulz-Narres  writes:

> Hello,
>
> I was wondering wether amanda traffic between client and server is
> transport encrypted.
>
> From the mailing list archive and Documentation I assume, that it is
> only encrypted if I use SSH as transport. Is that correct?
>

Or something like stunnel. Yes.
KJ


-- 
http://wolnelektury.pl/wesprzyj/teraz/

Re: How to "unlable" a tape

[Kamil Jońca]

Olivier  writes:

> Hi,
>
> I have been trying to find the command opposite to amlabel: completely
> remove a tape from Amanda.
>
> I know there is amadmin no-reuse, but suppose the tape had been
> completely destroyed and is not readable anymore, there should be a way
> to tell Amanda it should completely forget about that tape, remove any
> index it can have aboutt he tape.
>
> What would be the command then?

man amrmtape

KJ


-- 
http://wolnelektury.pl/wesprzyj/teraz/
You will be married within a year.

Re: Weird amdump behavior

[Kamil Jońca]

Nathan Stratton Treadway  writes:

> On Fri, Sep 20, 2019 at 00:17:05 -0400, Nathan Stratton Treadway wrote:
>> On Sat, Aug 04, 2018 at 08:51:38 +0200, Kamil Jo?ca wrote:
>> > PS. I have simple created disklist file with 2 + 4 disk which can
>> > reproduce problem.
>
> KJ, did you ever figure out anything about what was causing your amdump
> to fail when run for multiple hosts?  Did you find a workaround?

No, and no.

KJ

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
No sooner had Edger Allen Poe
Finished his old Raven,
then he started his Old Crow.

Re: amandad defunct on client calcsize/sendsize hung

[Kamil Jońca]

Robert Reilly  writes:

> All, Is there a way to run the backup manually to see whats happening on the 
> client outside of running strace on amdump ?
> Rob

Please:
 1. clear environment (kill zombies, hung processes, etc)
 2. try to backup single host
 3. try to backup all hosts EXCEPT amanda server
I am curious if it is my problem (see:
 https://www.mail-archive.com/amanda-users@amanda.org/msg49744.html)

KJ

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
There are no manifestos like cannon and musketry.
-- The Duke of Wellington

Re: amandad defunct on client calcsize/sendsize hung

[Kamil Jońca]

Robert Reilly  writes:

> Hi, All I have setup Amanda, it works great for the Amanda server itself but 
> fails to backup any clients.
> Server:
> Ubuntu 18.04
> Amanda 3.5.1 ( ubuntu repo pkg )

Whats happening when you try to backup single client?

i.e. amdump  clienthost?

does it works?
KJ

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
Epperson's law:
When a man says it's a silly, childish game, it's probably
something his wife can beat him at.

Re: Weird amdump behavior

[Kamil Jońca]

"Austin S. Hemmelgarn"  writes:

> On 2018-07-30 00:38, Kamil Jońca wrote:
>> Gene Heskett  writes:
>>
>>> On Saturday 28 July 2018 08:30:27 Kamil Jońca wrote:
>>>
>>>> Gene Heskett  writes:
>>>>
>>>> [..]
>>>>
>>>>> Too many dumps per spindle, drive seeks take time=timeout?
>>
>> As I can see in gdb/strace planner hangs on "futex"
> 'futex' is short for 'Fast Userspace muTEX', it's a synchronization
> primitive.  Based on personal experience (not with Amanda, but just

I know that, and I suspect is that we have race condition somewhere.

> debugging software hangs in general), this usually means it's either a
> threading issue, or that you've ended up with a deadlock somewhere
> between processes.  Regardless, it's probably an issue on the local

agreed. :(

> system, and most likely only happens when backing up more than one
> client because you have more processes/threads involved and actually
> doing things in that case.
>
> This is probably going to sound stupid, but try
> updating/rebuilding/reinstalling Perl, whatever Perl packages Amanda

I have reinstalled all packages. No luck :(

KJ

PS. I have simple created disklist file with 2 + 4 disk which can
reproduce problem.

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
Garbage In -- Gospel Out.

Re: Weird amdump behavior

[Kamil Jońca]

"Stefan G. Weichinger"  writes:

> Am 28.07.2018 um 14:30 schrieb Kamil Jońca:
>> Gene Heskett  writes:
>>
>> [..]
>>
>>>
>>> Too many dumps per spindle, drive seeks take time=timeout?
>> 1. I do not configure spindle at all.
>> 2. I think [1] I can reproduce this errror with only one disk per host
>> KJ
>> [1] It is quite time consuming
>
> As Gene suggested, consider setting/using spindles. Won't hurt.
I enumerated dles, and now each have its own spindle (but I think, that
I should change maxdumps too).
So far - no change - amanda hangs

> Maybe show us your actual config and the report emails.

I sent it in one of my previous mails. I can try to prepare minimal
hanging config (but it can take a while.)

KJ

-- 
http://wolnelektury.pl/wesprzyj/teraz/
Blame Saint Andreas -- it's all his fault.

Re: Weird amdump behavior

[Kamil Jońca]

Gene Heskett  writes:

> On Saturday 28 July 2018 08:30:27 Kamil Jońca wrote:
>
>> Gene Heskett  writes:
>>
>> [..]
>>
>> > Too many dumps per spindle, drive seeks take time=timeout?

As I can see in gdb/strace planner hangs on "futex"

>>
>> 1. I do not configure spindle at all.
>
> So its possible to have multiple dumps from the same spindle at the same 
> time.

No. There is another parameter,
--8<---cut here---start->8---
 maxdumps int
   Default: 1. The maximum number of backups from a single host
   that Amanda will attempt to run in parallel. See also the
   inparallel option.
--8<---cut here---end--->8---

And I use default value, so I have at most one dump per host at once
(and I am quite happy with this)

Of course I can change spindles for testing, but, to be honest, I do
not understand, how should that help.


>
> Please, give every disk in each machine its own unique spindle number. 
> Your backups should be done much faster.

I do not want faster dumps . I want working dumps. 

KJ

-- 
http://wolnelektury.pl/wesprzyj/teraz/
Reichel's Law:
A body on vacation tends to remain on vacation unless acted upon by
an outside force.

Re: Weird amdump behavior

[Kamil Jońca]

Gene Heskett  writes:

[..]

>
> Too many dumps per spindle, drive seeks take time=timeout?
1. I do not configure spindle at all.
2. I think [1] I can reproduce this errror with only one disk per host
KJ
[1] It is quite time consuming

-- 
http://wolnelektury.pl/wesprzyj/teraz/
Nie przerywaj mi, kiedy ja przerywam
--W.Churchill

Re: Weird amdump behavior

[Kamil Jońca]

"Stefan G. Weichinger"  writes:

> Am 08.07.2018 um 09:38 schrieb Kamil Jońca:
>> "Stefan G. Weichinger"  writes:
>>> amdump succeeds? Ok, then this might be your issue.
>>> I understood amcheck AND amdump fails.
>> As I wrote earlier:
>> 1. amdump WORKS when is called for ONE host, but
>> 2. almost always FAILS when is called for more than one host
>
> That should not happen, never saw this myself.
> Which release of amanda, which OS?
amdump-3.5.1
on debian


> That is just an idea, I don't know if that might be your issue, just
> one step I would take.

Well, but should it fail regardless number of hosts?
KJ

-- 
http://wolnelektury.pl/wesprzyj/teraz/
Software is like sex; it's better when it's free.
-- Linus Torvalds

Re: Weird amdump behavior

[Kamil Jońca]

Gene Heskett  writes:

[...]
>>
>> As I wrote earlier:
>> 1. amdump WORKS when is called for ONE host, but
>> 2. almost always FAILS when is called for more than one host
>>
>>
>> KJ
>
> You may want to check on spindle number in the amanda.conf. Generally you 
> should not have 2 invocations running against the same spindle(physical 
> disk) at the same time.
>

I did not touch spindles.
KJ

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
If a group of _N persons implements a COBOL compiler, there will be _N-1
passes.  Someone in the group has to be the manager.
-- T. Cheatham

Re: Weird amdump behavior

[Kamil Jońca]

"Stefan G. Weichinger"  writes:

> Am 08.07.2018 um 09:16 schrieb Kamil Jońca:
>> "Stefan G. Weichinger"  writes:
>>>> HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
>>>> /home/maria/Backup/outlook/provident 
>>>> (/home/maria/Backup/outlook/provident): Permission denied
>>>
>>> Sounds to me as if your user "backup" isn't member of group disk or so.
>>
>> I think this is the case.
>>
>> https://wiki.zmanda.com/index.php/Why_does_amcheck_fail_while_amdump_succeeds%3F
>
> amdump succeeds? Ok, then this might be your issue.
> I understood amcheck AND amdump fails.
As I wrote earlier:
1. amdump WORKS when is called for ONE host, but
2. almost always FAILS when is called for more than one host


KJ

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
Any time things appear to be going better, you have overlooked something.

Re: Weird amdump behavior

[Kamil Jońca]

"Stefan G. Weichinger"  writes:

> Am 07.07.2018 um 20:13 schrieb Kamil Jońca:
>
>> Amanda Backup Client Hosts Check
>> 
>> HOST alfa.kjonca ERROR: TMPDIR '/tmp/ampgsql' is NOT an acessible
>> directory
IIRC this is needed with recovery, not dump.

>
> Doesn't exist? Is it a DLE? (not a big issue)
>
>> HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
>> /home/maria/Backup/outlook/provident (/home/maria/Backup/outlook/provident): 
>> Permission denied
>
> Sounds to me as if your user "backup" isn't member of group disk or so.
>


I think this is the case.

https://wiki.zmanda.com/index.php/Why_does_amcheck_fail_while_amdump_succeeds%3F


>

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
There's another way to survive.  Mutual trust -- and help.
-- Kirk, "Day of the Dove", stardate unknown

Re: Weird amdump behavior

[Kamil Jońca]

"Stefan G. Weichinger"  writes:

>> 
>> $sudo -u backup amcheck alfa
>> Amanda Tape Server Host Check
>> -
>> NOTE: Holding disk '/var/spool/amanda/holding': 212353024 KB disk space 
>> available, using 157286400 KB as requested
>> Searching for label 'alfa-3-24':Label 'alfa-3-24' not found
>> slot 51: volume 'alfa-6-51'
>> Will write to volume 'alfa-6-51' in slot 51.
>> NOTE: skipping tape-writable test
>> umount: /var/spool/amanda/vtapes: umount failed: Operation not permitted.
>
> Seems as if your user backup is not allowed to mount/unmount that
> tape-directory. Either allow via flag "user" in fstab or change the
> properties MOUNT and UMOUNT to "no" in your amanda.conf
>
>> NOTE: index dir '/var/lib/amanda/alfa/state/index/alfa.kjonca/pg-10' does 
>> not exist
>>   it will be created on the next run
>> [... ]
>> 
>> Server check took 0.657 seconds
>> Amanda Backup Client Hosts Check
>> 
>> ERROR: alfa.kjonca: selfcheck request failed: timeout waiting for ACK
>> Client check: 2 hosts checked in 34.501 seconds.  1 problem found.
>
> This is your main issue.
>

And:
sudo -u backup amcheck alfa alfa.kjonca
Amanda Tape Server Host Check
-
NOTE: Holding disk '/var/spool/amanda/holding': 210120704 KB disk space 
available, using 157286400 KB as requested
Searching for label 'alfa-3-24':Label 'alfa-3-24' not found
slot 67: volume 'alfa-6-67'
Will write to volume 'alfa-6-67' in slot 67.
NOTE: skipping tape-writable test
NOTE: index dir '/var/lib/amanda/alfa/state/index/alfa.kjonca/pg-10' does not 
exist
  it will be created on the next run
[...]
Server check took 0.794 seconds
Amanda Backup Client Hosts Check

HOST alfa.kjonca ERROR: TMPDIR '/tmp/ampgsql' is NOT an acessible directory
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/maria/Backup/outlook/provident (/home/maria/Backup/outlook/provident): 
Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/maria/Backup/outlook/bosh/3 (/home/maria/Backup/outlook/bosh/3): 
Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/maria/Backup/outlook/bosh/2 (/home/maria/Backup/outlook/bosh/2): 
Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/maria/Backup/outlook/bosh/1 (/home/maria/Backup/outlook/bosh/1): 
Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/maria/Backup/outlook/bosh (/home/maria/Backup/outlook/bosh): Permission 
denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/maria/Backup/outlook (/home/maria/Backup/outlook): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access /home/kjonca/Mail 
(/home/kjonca/Mail): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access /home/kjonca/VM 
(/home/kjonca/VM): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/kjonca/Downloads/youtube (/home/kjonca/Downloads/youtube): Permission 
denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/kjonca/Downloads/czasopisma (/home/kjonca/Downloads/czasopisma): 
Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/kjonca/Downloads/chomik.pl (/home/kjonca/Downloads/chomik.pl): Permission 
denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/kjonca/Downloads/2018 (/home/kjonca/Downloads/2018): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/kjonca/Downloads/2017 (/home/kjonca/Downloads/2017): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/kjonca/Downloads/2016 (/home/kjonca/Downloads/2016): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/kjonca/Downloads/2015 (/home/kjonca/Downloads/2015): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/kjonca/Downloads/2014 (/home/kjonca/Downloads/2014): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/kjonca/Downloads/2013 (/home/kjonca/Downloads/2013): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access /home/kjonca/Downloads 
(/home/kjonca/Downloads): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/kjonca/CalibreLibrary (/home/kjonca/CalibreLibrary): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access /home/kjonca/Muzyka 
(/home/kjonca/Muzyka): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access 
/home/kjonca/archiwum/obrazy (/home/kjonca/archiwum/obrazy): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access /home/kjonca/wd/kw/dane 
(/home/kjonca/wd/kw/dane): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access /home/kjonca/archiwum 
(/home/kjonca/archiwum): Permission denied
HOST alfa.kjonca ERROR: alfa.kjonca: Could not access /home/kjonca/backup 
(/home/kjonca/backup): Permission denied

Re: Weird amdump behavior

[Kamil Jońca]

"Stefan G. Weichinger"  writes:

> Am 2018-07-07 um 14:34 schrieb Kamil Jońca:
>
>> under this link is amanda.tgz 
>> https://drive.google.com/open?id=1QObymK4VWV3U-t81UjjQAK2AUt236pyi
>> 
>> which contains
>> amanda.conf
>> disklist
>> report email
>
> [..]
>
>> It seems to have the same problem ... 
>> 
>> $sudo -u backup amcheck alfa
>> Amanda Tape Server Host Check
>> -
>> NOTE: Holding disk '/var/spool/amanda/holding': 212353024 KB disk space 
>> available, using 157286400 KB as requested
>> Searching for label 'alfa-3-24':Label 'alfa-3-24' not found
>> slot 51: volume 'alfa-6-51'
>> Will write to volume 'alfa-6-51' in slot 51.
>> NOTE: skipping tape-writable test
>> umount: /var/spool/amanda/vtapes: umount failed: Operation not permitted.
>
> Seems as if your user backup is not allowed to mount/unmount that
> tape-directory. Either allow via flag "user" in fstab or change the
> properties MOUNT and UMOUNT to "no" in your amanda.conf

Usually I use "--no-taper" option with amdump, so it is not the problem
here.


>
>> NOTE: index dir '/var/lib/amanda/alfa/state/index/alfa.kjonca/pg-10' does 
>> not exist
>>   it will be created on the next run
>> [... ]
>> 
>> Server check took 0.657 seconds
>> Amanda Backup Client Hosts Check
>> 
>> ERROR: alfa.kjonca: selfcheck request failed: timeout waiting for ACK
>> Client check: 2 hosts checked in 34.501 seconds.  1 problem found.
>
> This is your main issue.
>
> Work through
>
> http://wiki.zmanda.com/index.php/Selfcheck_request_failed
>
> and specifically get your xinetd right / or your amanda service if you
> use systemd:
>
> http://wiki.zmanda.com/index.php/How_To:Configure_bsdtcp_authentication
>
> You try "auth local" ... manpage says:
>
> "The authenticated peer hostname for this authentication will be as
> reported by gethostname if that succeeds, or the fixed string localhost
> otherwise.
>
> This method requires the host name in the DLE to be one that refers to
> this host. Any name for which gataddrinfo is able to return at least one
> address usable as the local end of a socket will work."
>
> So check if your hostname resolution brings back "alfa.kjonca"
> correctly.
>
> Personally I'd prefer bsdtcp-auth, mainly because it works for me in
> dozens of setups.
>

1. SO far my setup worked for me for 4 years. Something happen with latests
amanda releases.
2. When I try amdump alfa.kjonca (only one host) everything works ok 
3. When I try amdump all hosts - effects are as above, but ...
4. rarely (very rarely) amdump for all hosts works as expected
5. I tried to use both bsdtcp and local auth - no change.
6. dig alfa.kjonca/host alfa.kjonca works as expected

KJ

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
Price does not include taxes.

Re: Weird amdump behavior

[Kamil Jońca]

"Stefan G. Weichinger"  writes:

> Am 04.07.2018 um 19:47 schrieb Kamil Jońca:
>> "Stefan G. Weichinger"  writes:
>>> Do not use "localhost" in DLEs. It's mentioned in the docs as well
>>
>> I must overlooked it. But would you point me to the right manual?
>> in man disklist/amanda.conf/amanda-match I cannot found nothing useful.
>
> Hm, can't find it right now ;-) I remember I wrote a howto years ago
> mentioning this, maybe it has been removed.
>
> Anyway.
>
>>> (although more of a recover-issue than a backup-issue).
>>>
>>> Use FQDNs, check your /etc/hosts and DNS then planner should work fine.
>>>
>> I changed 'localhost' to fully qualified name and no succcess  :(
>
> Provide your amanda.conf, disklist and the output of the commands you
> run and/or a report email.
>
under this link is amanda.tgz 
https://drive.google.com/open?id=1QObymK4VWV3U-t81UjjQAK2AUt236pyi

which contains
amanda.conf
disklist
report email


> "no success" is no useful start for helping you.
>
> Or for a start:
>
> does amcheck work correctly?
>

It seems to have the same problem ... 

$sudo -u backup amcheck alfa
Amanda Tape Server Host Check
-
NOTE: Holding disk '/var/spool/amanda/holding': 212353024 KB disk space 
available, using 157286400 KB as requested
Searching for label 'alfa-3-24':Label 'alfa-3-24' not found
slot 51: volume 'alfa-6-51'
Will write to volume 'alfa-6-51' in slot 51.
NOTE: skipping tape-writable test
umount: /var/spool/amanda/vtapes: umount failed: Operation not permitted.

NOTE: index dir '/var/lib/amanda/alfa/state/index/alfa.kjonca/pg-10' does not 
exist
  it will be created on the next run
[... ]

Server check took 0.657 seconds
Amanda Backup Client Hosts Check

ERROR: alfa.kjonca: selfcheck request failed: timeout waiting for ACK
Client check: 2 hosts checked in 34.501 seconds.  1 problem found.
(brought to you by Amanda 3.5.1)

-- 
http://wolnelektury.pl/wesprzyj/teraz/
"Truth never comes into the world but like a bastard, to the ignominy
of him that brought her birth."
-- Milton

Re: Weird amdump behavior

[Kamil Jońca]

"Stefan G. Weichinger"  writes:

> Am 2018-07-02 um 20:12 schrieb Kamil Jońca:
>> Debra S Baddorf  writes:
>> 
>>> Are you using “localhost”  as a node name?   Perhaps you could try 
>>> using a “fully qualified domain name” instead, in your disklist file
>>> and everywhere.  I’ve heard issues with “localhost”  in the past.
>
>> It is worse I think.
>> 
>> I make some calls to amdump recently, and some of these calls ended
>> sucessfully.
>> 
>> So we have two scenarios:
>> 1. amdump ends with success
>> 2. amdump ends with error, adn all localhost DLEs have estimate failed.
>> Another thing: when amdump fails there is hanging alone planner
>> process.
>> 
>> It seems like race condition here.
>
> Do not use "localhost" in DLEs. It's mentioned in the docs as well

I must overlooked it. But would you point me to the right manual?
in man disklist/amanda.conf/amanda-match I cannot found nothing useful.

> (although more of a recover-issue than a backup-issue).
>
> Use FQDNs, check your /etc/hosts and DNS then planner should work fine.
>
I changed 'localhost' to fully qualified name and no succcess  :(
Kj

-- 
http://wolnelektury.pl/wesprzyj/teraz/
A man's house is his hassle.

Re: Weird amdump behavior

[Kamil Jońca]

Debra S Baddorf  writes:

> Are you using “localhost”  as a node name?   Perhaps you could try 
> using a “fully qualified domain name” instead, in your disklist file
> and everywhere.  I’ve heard issues with “localhost”  in the past.
>
> Deb Baddorf
> Fermilab
>
>
>> On Jul 1, 2018, at 9:59 AM, Kamil Jońca  wrote:
>> 
>> kjo...@o2.pl (Kamil Jońca) writes:
>> 
>>> Recently I noticed weird behavior of amdump.
>>> 
>>> I have several localhost DLEs.
>>> 
>>> When I call
>>> 
>>> sudo -u backup amdump --no-taper alfa localhost
>>> 
>>> everything goes ok.
>>> also 
>>> 
>>> sudo -u backup amdump --no-taper alfa laptop.kjonca
>>> 
>>> works as expected and backup my laptop
>>> 
>>> but when I call
>>> 
>>> sudo -u backup amdump --no-taper alfa
>>> 
>>> I got
>>> ERROR planner Request to localhost failed: timeout waiting for ACK
>>> 
>>> What should I check/change?
>>> KJ


It is worse I think.

I make some calls to amdump recently, and some of these calls ended
sucessfully.

So we have two scenarios:
1. amdump ends with success
2. amdump ends with error, adn all localhost DLEs have estimate failed.
Another thing: when amdump fails there is hanging alone planner
process.

It seems like race condition here.
KJ

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
Unprovided with original learning, unformed in the habits of thinking,
unskilled in the arts of composition, I resolved to write a book.
-- Edward Gibbon

Re: Weird amdump behavior

[Kamil Jońca]

kjo...@o2.pl (Kamil Jońca) writes:

> Recently I noticed weird behavior of amdump.
>
> I have several localhost DLEs.
>
> When I call
>
> sudo -u backup amdump --no-taper alfa localhost
>
> everything goes ok.
> also 
>
> sudo -u backup amdump --no-taper alfa laptop.kjonca
>
> works as expected and backup my laptop
>
> but when I call
>
> sudo -u backup amdump --no-taper alfa
>
> I got
> ERROR planner Request to localhost failed: timeout waiting for ACK
>
> What should I check/change?
> KJ

Unfortunately this issue still persists.  Any hints?
KJ
-- 
http://wolnelektury.pl/wesprzyj/teraz/
It was pity stayed his hand.
"Pity I don't have any more bullets," thought Frito.
-- _Bored_of_the_Rings_, a Harvard Lampoon parody of Tolkein

Weird amdump behavior

[Kamil Jońca]

Recently I noticed weird behavior of amdump.

I have several localohost DLEs.

When I call
--8<---cut here---start->8---
sudo -u backup amdump --no-taper alfa localhost
--8<---cut here---end--->8---
ewerything goes ok.
also 
--8<---cut here---start->8---
sudo -u backup amdump --no-taper alfa laptop.kjonca
--8<---cut here---end--->8---
works as expected and backup my laptop

but when I call

--8<---cut here---start->8---
sudo -u backup amdump --no-taper alfa
--8<---cut here---end--->8---

I got
ERROR planner Request to localhost failed: timeout waiting for ACK

What should I check/change?
KJ



-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
Television has proved that people will look at anything rather than each other.
-- Ann Landers

Re: amanda 3.5 (debian) - multiple instances - is it normal?

[Kamil Jońca]

Jean-Louis Martineau  writes:

> amanda-3.5 allow to run multiple backup in parallel.
> This to allow to run different amdump on different host/dle at the same time 
> and to allow client initiated backup to run at the same time.
>
> Jean-Louis

Yes, I have read about it, sorry for the noise.
I have another questions.
How can I list currently processed backups?
How can I pass backup to amcleanup?
KJ
-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
Only through hard work and perseverance can one truly suffer.

amanda 3.5 (debian) - multiple instances - is it normal?

[Kamil Jońca]

So far amanda was unable to run multiple instances simultaneously.
But recently I saw that exists two amdump processes dumping the same
disk (although to different holiding disk). This is normal?
KJ

-- 
http://wolnelektury.pl/wesprzyj/teraz/
To kick or not to kick...
-- Somewhere on IRC, inspired by Shakespeare

"taperflush criteria not met"

[Kamil Jońca]

I have problem with flushing - recently almost alway amflush does not
 flush complete dump but write "PARTIAL" in report.

and moreover i have message like this.

 localhost /srv/samba lev 0  partial taper: taperflush criteria not met

taperflush parameters are not touched.

--8<---cut here---start->8---
#amflush --version 
amflush-3.3.8
--8<---cut here---end--->8---
its amanda packaged by debian.

What should I check?

KJ
-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
We're living in a golden age.  All you need is gold.
-- D. W. Robertson.

amlabel problem

[Kamil Jońca]

I want to add some new virtual tapes, but I got:
--8<---cut here---start->8---
Reading label...
Found an empty tape.

amlabel: Can't use an undefined value as a subroutine reference at 
/usr/sbin/amlabel line 120.
--8<---cut here---end--->8---

lines from sbin/amlabel:

--8<---cut here---start->8---
117 if ($res) {
118 $res->release(finished_cb => sub {
119 # ignore error
120 $finished_cb->()
121 });
--8<---cut here---end--->8---
KJ

-- 
http://wolnelektury.pl/wesprzyj/teraz/
No line available at 300 baud.

Re: amrmtape - what I'm doing wrong?

[Kamil Jońca]

Jean-Louis Martineau martin...@zmanda.com writes:

 Look for error in the debug file.

Ahh.

--8---cut here---start-8---
Sun Jun 21 14:42:19 2015: thd-0x1829e00: amrmtape: new Amanda::Changer::Error: 
type='fatal', message='slot 1 doesn't exists '/var/spool/amanda/vtapes/slot1''
--8---cut here---end---8---

After removing tape I delete slot dir, and after first tape there were
no slot for vtapes.

(I have in amanda.conf:
define changer removable-disk {
tpchanger chg-disk:/var/spool/amanda/vtapes
property NUM-SLOT 32
property AUTO-CREATE-SLOT no
[...]
}

Thank you.
KJ

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
Today is the tomorrow you worried about yesterday.

Re: ampgsql - ignores amanda-client.conf ?

[Kamil Jońca]

Jean-Louis Martineau martin...@zmanda.com writes:

 man ampgsql

   STATEDIR
   Directory for saving state about backups already made. The default
   is set when Amanda is built by the --with-gnutar-listdir configure
   option.


Thanks. Now works like a charm.
KJ

-- 
http://stopstopnop.pl/stop_stopnop.pl_o_nas.html
Clear the laundromat!!  This whirl-o-matic just had a nuclear meltdown!!

amrmtape - what I'm doing wrong?

[Kamil Jońca]

When I try to execute:
--8---cut here---start-8---
sudo -u backup amrmtape --cleanup --erase -v alfa alfa-6-100
--8---cut here---end---8---
I got:
--8---cut here---start-8---
amrmtape: Can't locate object method load via package 
Amanda::Changer::Error at /usr/sbin/amrmtape line 299.
--8---cut here---end---8---

Amanda 3.3.6 packaged by debian
KJ
-- 
http://wolnelektury.pl/wesprzyj/teraz/
EMACS = Emacs Manuals Are Cryptic and Surreal

ampgsql - ignores amanda-client.conf ?

[Kamil Jońca]

I have to backup some postgresql databases. I used to use ampgsql for
that.
Some days ago I had to make full backup. And since then I have full
backup of postgres every day.
My suspection is that ampgsql ignores amanda-client.conf settings

my amanda-client.conf
--8---cut here---start-8---
gnutar_list_dir /var/lib/amanda/alfa/gnutar-lists
amandates /var/lib/amanda/alfa/amandates
--8---cut here---end---8---

exceprts from /var/log/amanda/client/alfa/ampgsql.20150604151849000.debug

--8---cut here---start-8---
Thu Jun  4 19:15:11 2015: thd-0x2baea00: ampgsql: size of generated tar file: 
16783360
Thu Jun  4 19:15:11 2015: thd-0x2baea00: ampgsql: statefile: 
/var/lib/amanda/gnutar-lists/ampgsql-%2fvar%2frun%2fpostgresql%2f-pg%2d9%2e4-5
Thu Jun  4 19:15:11 2015: thd-0x2baea00: ampgsql: Failed to write state file
--8---cut here---end---8---

Moreover when changed state file directory to its default location, then
ampgsq function properly again.
KJ

-- 
http://wolnelektury.pl/wesprzyj/teraz/
Yow!  Am I having fun yet?

Re: ampgsql - some questions

[Kamil Jońca]

Jean-Louis Martineau martin...@zmanda.com writes:

[...]


 This create two properties of the dumptype, you want application properties:
 It is a bug to use the 'PG-' prefix on the application properties.
 The 'PG-' prefix must be use only if you define the properties in
 amanda-client.conf, don't do it, it is deprecated.

Yes, I know it's deprecated (as written
http://wiki.zmanda.com/man/ampgsql.8.html), but I think, that manual is
not clear here about PG- prefix, moreover I can't see DATADIR property
for application. 

[...]


 define dumptype ampgsql-local {
 gnutar-local  # You might need this if you're setting 
 'auth bsdtcp' in your global dumptype settings, for example
 program APPLICATION
 application {
app-ampgsql
property ARCHIVEDIR /var/lib/postgresql/archive  # not required, 
 it is set in app-ampgsql
property DATADIR /var/lib/postgresql/9.1/main# you could set 
 it in app-ampgsql
 }
 }



 Either the diskdevice in disklist or the DATADIR application property.

Thanks for your reply.
BTW. I thought, that application parameter can accept only name of
previously defined application.


--8---cut here---start-8---
define dumptype ampgsql-local {
 program APPLICATION
 application {
app-ampgsql
property ARCHIVEDIR /var/lib/postgresql/archive  # not required, it 
is set in app-ampgsql
property DATADIR /var/lib/postgresql/9.1/main# you could set it 
in app-ampgsql
 }
}
--8---cut here---end---8---

KJ



-- 
http://blogdebart.pl/2012/06/24/hiena/
I have an existential map.  It has You are here written all over it.
-- Steven Wright

ampgsql - some questions

[Kamil Jońca]

I tried to use ampgsql to backup mu personal database cluster (over a
dozen GBs) finally I think, I did it, but there is some ambiguity for
me.
I used 
http://wiki.zmanda.com/man/ampgsql.8.html
http://wiki.zmanda.com/index.php/How_To:Use_Amanda_to_Back_Up_PostgreSQL
as a reference. (I used debian's 1:3.3.3-1 version)

My first config (excerpts)
[...]
amanda.conf
--8---cut here---start-8---
define dumptype gnutar {
program GNUTAR
}
define dumptype gnutar-local {
gnutar
auth local
}

define application-tool app-ampgsql {
comment ampgsql
plugin ampgsql
property ARCHIVEDIR /var/lib/postgresql/archive
property TMPDIR /tmp/ampgsql
property DIRECTORY /tmp/amanda-pgsql/
property HOST /var/run/postgresql/
property USER backup
}

define dumptype ampgsql-local {
gnutar-local  # You might need this if you're setting 'auth 
bsdtcp' in your global dumptype settings, for example
  
program APPLICATION
application app-ampgsql
property PG-ARCHIVEDIR /var/lib/postgresql/archive
property PG-DATADIR /var/lib/postgresql/9.1/main

}
--8---cut here---end---8---


disklist
--8---cut here---start-8---
localhost postgres  ampgsql-local
--8---cut here---end---8---

(example from howto was  foo.example.com bar dt_ampgsql)

But this did not work :(

After read some debug and log files I found that tar called from ampgsql
tries to backup postrgresql dir which was clearly wrong. 
then I changed disklist to 

--8---cut here---start-8---
localhost postgres /var/lib/postgresql/9.1/main ampgsql-local
--8---cut here---end---8---

And this seems to work. (At least dump contents looks reasonable)
But 
1. Am I misssed something about cluster data dir in dikslist?
2. If I have to use diskdevice in disklist entry - what for is PG-DATADIR?

KJ

-- 
http://blogdebart.pl/2010/03/17/dalsze-przygody-swinki-w-new-jersey/
World War Three can be averted by adherence to a strictly enforced dress code!

Re: full dump in degraded mode - possible?

[Kamil Jońca]

C.Scheeder christ...@scheeder.de writes:

 Am 29.03.2013 06:30, schrieb Kamil Jońca:
 As far I know amanda does not do full dump when no tape (vtape in my
 case) is not available.
 Is there any way to force it?
 If not what is the rationale for this?

 (I would prefer keep disk with vtapes as short as possible turned on)
 KJ


 Hm,

 what about RTFM ?
Shame on me :( 

 ;-)

 especialy man amanda.conf,
 where in the section about global parameters, it tells you
 something about the parameter reserve int, which defaults to 100,

Thanks. 

KJ


-- 
http://modnebzdury.wordpress.com/2009/10/01/niewiarygodny-list-prof-majewskiej-wprowadzenie/
Political history is far too criminal a subject to be a fit thing to
teach children.
-- W. H. Auden

Re: flush-threshold-* parameters - some clarification needed.

[Kamil Jońca]

Jean-Louis Martineau martin...@zmanda.com writes:

 On 03/27/2013 06:39 AM, Kamil Jońca wrote:
 (This is duplicate mail, previously sent, but I weren't subscribed)
 In some sites there is suggestion, that config:
 --8---cut here---start-8---
 #flush-threshold-dumped 100 # (or more)
 #flush-threshold-scheduled 100 # (or more)
 #taperflush 100
 #autoflush yes
 --8---cut here---end---8---
 helps to do better tape utilization.

 But how behave amanda in my configuration (+ these four lines above):

 I have defined 1gb vtapes, placed onto four hard disks rotated
 (ie. first flush go to disk 1, second to 2, third to 3, fourth to 4,
 fifth to 1, and so on)

 say on first disk are tapes: alfa-1-1 to alfa-1-10
 on second:  alfa-2-1 to alfa-2-10
 on third: alfa-3-1 to alfa-3-10
 on fourth: alfa-4-1 to alfa-4-10

 One day dumps takes 1,5GB and goes to tape alfa-1-1 and (partially)
 alfa-1-2.
 Amanda should not write to alfa-1-2 if the above setting are commented out.
 Are you using a holding a disk? It must be larger than a vtape (1gb).


My bad. I forgot to say that I have  runtapes greater than 1.
And my question is what amanda do if above flush settings are not
commmented.

KJ


-- 
http://blogdebart.pl/2009/12/22/mamy-chorych-dzieci/
Wiesz, tryb tekstowy w Linuksie ma się tak do DOSu
jak F-117A do paralotni. (c) Dawid Kuroczko

Re: flush-threshold-* parameters - some clarification needed.

[Kamil Jońca]

Jean-Louis Martineau martin...@zmanda.com writes:

 On 03/27/2013 08:58 AM, Kamil Jońca wrote:
 Jean-Louis Martineau martin...@zmanda.com writes:

 On 03/27/2013 06:39 AM, Kamil Jońca wrote:
 (This is duplicate mail, previously sent, but I weren't subscribed)
 In some sites there is suggestion, that config:
 --8---cut here---start-8---
 #flush-threshold-dumped 100 # (or more)
 #flush-threshold-scheduled 100 # (or more)
 #taperflush 100
 #autoflush yes
 --8---cut here---end---8---
 helps to do better tape utilization.

 But how behave amanda in my configuration (+ these four lines above):

 I have defined 1gb vtapes, placed onto four hard disks rotated
 (ie. first flush go to disk 1, second to 2, third to 3, fourth to 4,
 fifth to 1, and so on)

 say on first disk are tapes: alfa-1-1 to alfa-1-10
 on second:  alfa-2-1 to alfa-2-10
 on third: alfa-3-1 to alfa-3-10
 on fourth: alfa-4-1 to alfa-4-10

 One day dumps takes 1,5GB and goes to tape alfa-1-1 and (partially)
 alfa-1-2.
 Amanda should not write to alfa-1-2 if the above setting are commented out.
 Are you using a holding a disk? It must be larger than a vtape (1gb).

 My bad. I forgot to say that I have  runtapes greater than 1.
 And my question is what amanda do if above flush settings are not
 commmented.

 Amanda should not write to alfa-1-2, the backup should stay in the
 holding disk.
Yes, I use holding disk.

So far I do not have these lines and amanda writes backup to disk (for
example today backup below):
--8---cut here---start-8---
-rw-r- 1 backup backup243 Mar 27 16:27 
/var/spool/amanda/vtapes/state
-rw--- 1 backup backup  39436 Mar 27 16:23 
/var/spool/amanda/vtapes/slot8/8.localhost._srv_samba_kopie.2
-rw--- 1 backup backup  33328 Mar 27 16:23 
/var/spool/amanda/vtapes/slot8/9.localhost._var_lib_samba.1
-rw--- 1 backup backup   25797104 Mar 27 16:23 
/var/spool/amanda/vtapes/slot8/00010.localhost._var_spool_news_archive.1
-rw--- 1 backup backup   58407680 Mar 27 16:23 
/var/spool/amanda/vtapes/slot8/3.localhost._home_kjonca_Mail.6
-rw--- 1 backup backup  63076 Mar 27 16:23 
/var/spool/amanda/vtapes/slot8/4.localhost._home_kjonca_Muzyka.1
-rw--- 1 backup backup   16056464 Mar 27 16:23 
/var/spool/amanda/vtapes/slot8/5.localhost._home_kjonca_archiwum.2
-rw--- 1 backup backup   48955192 Mar 27 16:23 
/var/spool/amanda/vtapes/slot8/6.localhost._srv_archiwa.2
-rw--- 1 backup backup  43736 Mar 27 16:23 
/var/spool/amanda/vtapes/slot8/7.localhost._srv_samba.1
-rw--- 1 backup backup  34024 Mar 27 16:23 
/var/spool/amanda/vtapes/slot8/2.localhost._home_kjonca_Drivers.1
-rw--- 1 backup backup  133441876 Mar 27 16:23 
/var/spool/amanda/vtapes/slot8/1.localhost._home_kjonca_Downloads.2
-rw--- 1 backup backup  32768 Mar 27 16:23 
/var/spool/amanda/vtapes/slot8/0.alfa-2-8
-rw--- 1 backup backup  965672960 Mar 27 16:23 
/var/spool/amanda/vtapes/slot7/8.localhost._home_kjonca_Downloads.2
-rw--- 1 backup backup   53746648 Mar 27 16:23 
/var/spool/amanda/vtapes/slot7/6.localhost._home_kjonca.3
-rw--- 1 backup backup  81132 Mar 27 16:23 
/var/spool/amanda/vtapes/slot7/7.localhost._home_kjonca_CalibreLibrary.1
-rw--- 1 backup backup   34692104 Mar 27 16:23 
/var/spool/amanda/vtapes/slot7/2.bambus.kjonca._home.2
-rw--- 1 backup backup  59360 Mar 27 16:23 
/var/spool/amanda/vtapes/slot7/3.localhost._etc.1
-rw--- 1 backup backup 164232 Mar 27 16:23 
/var/spool/amanda/vtapes/slot7/4.localhost._home.1
-rw--- 1 backup backup   19114828 Mar 27 16:23 
/var/spool/amanda/vtapes/slot7/5.localhost._home_f4y.1
-rw--- 1 backup backup  52240 Mar 27 16:23 
/var/spool/amanda/vtapes/slot7/1.bambus.kjonca._etc.1
-rw--- 1 backup backup  32768 Mar 27 16:23 
/var/spool/amanda/vtapes/slot7/0.alfa-2-7
--8---cut here---end---8---

As we can see amanda fills two vtapes (slot7 and slot8). 
And my experience shows that (WITHOUT flush-* settings mentioned above)
at next run (with this disk) amanda will use next slot. Right?
What amanda do when these flush-* will be set to 100? 

 They should be written to vtape on the next run or when you have more than 
 1GB (100% of 1gb) of dump on holding disk.

 Uncommenting the above setting is useless if you do not use a holding
 disk.
Yes, of course. 

 Did you read the documentation? 
Yes. But I do not know which vtape amanda use at some of next run. 
(Because tomorrow I will put disk with labels alfa-3-*, on Friday alfa-4-*, 
on saturday alfa-1-*, and on Sunday again alfa-2-*)
Did you tried it?

I use amanda, but WITHOUT these settings.

KJ


-- 
http://blogdebart.pl/2012/06/24/hiena/
Try to get all of your posthumous medals in advance.

Re: flush-threshold-* parameters - some clarification needed.

[Kamil Jońca]

Jean-Louis Martineau martin...@zmanda.com writes:

 Kamil,

 The flush-threshold-* are useful for tape device, there is no
 advantage to use them with vtape.
 With a tape, if you don't fill it, there is unused space on it.
 With vtape, if you write 500MB to a 1GB vtape, then the vtape use
 500MB nothing more, the space is not lost, it can be use by another
 slot.

 I do not understand what you want to do.

Simple:
Say, I have 1 TB disk for vtapes.  
I can create 900 vtapes on this disk. But all of dumps have ~500MB, so
only half of disk are used. And only 900 days are remebered.
And I'm curious if I have better utilization for  disk, so I can have
longer period of dumps. 
For now, if I understand correctly, I can only do statistical estimate?


Another case:
I create vtapes which are +- DVD size, and it's rather important for me
not to have tapes with only some bytes on them.


In other post you said:

--8---cut here---start-8---
Not sure why you change the disk every days, but it's generally BAD
unless you do full every night.
--8---cut here---end---8---

Does this have something common with message at:
http://www.mail-archive.com/amanda-users@amanda.org/msg4.html ?

KJ


-- 
http://blogdebart.pl/2010/03/17/dalsze-przygody-swinki-w-new-jersey/
Uwolnić słonia !!!

Re: flush-threshold-* parameters - some clarification needed.

[Kamil Jońca]

Jean-Louis Martineau martin...@zmanda.com writes:

 On 03/27/2013 01:02 PM, Kamil Jońca wrote:
 Simple:
 Say, I have 1 TB disk for vtapes.
 I can create 900 vtapes on this disk. But all of dumps have ~500MB, so
 only half of disk are used. And only 900 days are remebered.
 And I'm curious if I have better utilization for  disk, so I can have
 longer period of dumps.

 Create 1800 vtapes.
 The number of slots should not directly depend on the length of the
 tapetype but on the used size of each slots.

But in this case there is a risk to get no space left on device when,
in turn, dumps get larger later.
What then?
KJ
-- 
http://sporothrix.wordpress.com/2011/01/16/usa-sie-krztusza-kto-nastepny/
Think sideways!
-- Ed De Bono