dibs-discussion — Mailing list for public discussion.

Re: [Dibs-discussion] DIBS status

[sophana <so...@zi...>]

Unfortunately, dibs isn't ready for production.
I haven't succeeded in making it working correctly.
Too bad because the concept is good.

Noah Diewald a écrit :
> I noticed that the list archives for DIBS end in 2006. I'm interested
> in using DIBS but I don't want to commit my organization to software
> that isn't being used or supported. Are there many people still using
> DIBS? Have people moved to some other solution?
>   

Re: [Dibs-discussion] DIBS status

[Jason M. <jhm...@to...>]

I haven't seen any activity for quite some time either; and last
I heard there were significant issues.  I think the project is
effectively dead.

-Jason Martin
On Mon, Jul 28, 2008 at 01:06:29PM -0600, Noah Diewald wrote:
> I noticed that the list archives for DIBS end in 2006. I'm interested
> in using DIBS but I don't want to commit my organization to software
> that isn't being used or supported. Are there many people still using
> DIBS? Have people moved to some other solution?
-- 
I haven't lost my mind -- it's backed up on tape somewhere.
This message is PGP/MIME signed.

[Dibs-discussion] DIBS status

[Noah D. <noa...@gm...>]

I noticed that the list archives for DIBS end in 2006. I'm interested
in using DIBS but I don't want to commit my organization to software
that isn't being used or supported. Are there many people still using
DIBS? Have people moved to some other solution?

[Dibs-discussion] dibs 0.92 looping.

[sophana <so...@zi...>]

Hi

I'm currently testing dibs, which looks like a very cool tool. The
concept is great, thanks for that.
Unfortunately it doesn't work very well.

I've setup 2 peers A and B, both with incoming port active.
I've exchanged gpg keys
in peer A, I added peer B
then did a manual backup of a directory.

Then I could see, that I forgot to add trust to the A gpg key to B

Whatever command I make from peer A, peer B takes a LOT of cpu (4minutes
of 2ghz cpu).

when I do a Ctrl C, it was in
Subject: <DIBS> DIBS ERROR

Command was '['/usr/bin/dibs.py', 'start_daemon']'.

Exception of type '<class exceptions.KeyboardInterrupt at 0xb7ee038c>':  ''.


Traceback:
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_ui.py", line 781,
in ?
    daemon = dibs_daemon.DIBSServer(dibs_options.daemonPort)
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_daemon.py", line
163, in __init__
    while (self.ProcessCommand() != 0):
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_daemon.py", line
262, in ProcessCommand
    return self.cmdTable[match.group('cmd')](match)
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_daemon.py", line
286, in ProcessGive
    match = self.ExpectResponse(rOKorI)
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_daemon.py", line
112, in ExpectResponse
    cmd = self.SocketReceive(dibs_constants.daemonCmdSize)
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_daemon.py", line
77, in SocketReceive
    except socket.error, (num, type):

Regards

Re: [Dibs-discussion] Status...

[J. W. H. <wr...@hu...>]

Ron Eringa wrote:
> 
> Hello Emin,
> 
> I just subscribed to the dibs mailing list, because I'm very interested
> in your dibs-initiative.
> Can you tell me something about the status of the current project?
> I saw a lot of developmont cycles in 2004, but only a few in 2005.
> Are your planning a new release or has the process run dead?
> 

You may also be interested in: http://allmydata.com/

Wren

[Dibs-discussion] Status...

[Ron E. <ron...@ph...>]

Hello Emin,

I just subscribed to the dibs mailing list, because I'm very interested in 
your dibs-initiative.
Can you tell me something about the status of the current project?
I saw a lot of developmont cycles in 2004, but only a few in 2005.
Are your planning a new release or has the process run dead?

Met vriendelijke groet/With kind regards, 

Ron Eringa
----------------------------------------------------------------------------------------------------------
Technical System Designer 
Philips TASS
Vonderweg 11   PO box 80060 Eindhoven (Netherlands)
Phone: +31-(0)40 27 43737 Fax: +31-(0)40 27 55419
----------------------------------------------------------------------------------------------------------

Re: [Dibs-discussion] fwd: dibs

[Emin M. <em...@al...>]

BrokenClock <bro...@fr...> writes:

> Your idea of making a night script is a good one, but.. it is not the
> way our computers run here, and I have no autority to change it for
> now. The computers have to be shutdown every evening. So it should be
> very useful for me to have a windows service for dibs...

OK.  I don't know how difficult it is to make a python program into a
windows service.  If anyone out there knows how to do this, that would
be a great help.  

> About the py2exe issue, I think that is should be cause by the fact
> you use dibsrc.py with import, and py2exe may tried to include
> it. Maybe you should be better to read an ini file, don't you think so?
> An idea?

Using an ini file instead of dibsrc.py would be easy to do, but I
think there are other problems with py2exe.  When I tried using py2exe
with your suggested setup file I kept getting an error about 
win32com.gen_py missing.  If you or someone else could email me a
setup file that allowed py2exe to work properly with a blank
dibsrc.py, then it wouldn't be hard to replace dibsrc.py with a
simpler initialization file.

In any case, I'll definitely make getting py2exe to work a priority
once I finish my current tasks and have time to do DIBS development
again.

Thanks,
-Emin

[Dibs-discussion] fwd: dibs

[BrokenClock <bro...@fr...>]

Hello everybody, hello Emin,

How I could have been unaware of this mailing list, I can't understand!=20
And SF is one of my favorite for a long time!
I'm pleased to subscribe it!

Your idea of making a night script is a good one, but.. it is not the=20
way our computers run here, and I have no autority to change it for now.=20
The computers have to be shutdown every evening. So it should be very=20
useful for me to have a windows service for dibs...
About the py2exe issue, I think that is should be cause by the fact you=20
use dibsrc.py with import, and py2exe may tried to include it. Maybe you=20
should be better to read an ini file, don't you think so?
An idea?

Many thanks and best regards,

BrokenClock


Emin Martinian a =E9crit :

>"BrokenClock" <bro...@fr...> writes:
>
> =20
>
>>* first, how DIBS discovers peers?
>>   =20
>>
>
>You can use the peer finder service described in the documentation at
>http://csua.berkeley.edu/~emin/source_code/dibs/dibs.html#Peer-Finder
>to advertise for peers or partner with peers who are already listed.
>There is currently no automated peering algorithm implemented, but
>this is on the list of things to do.
>
> =20
>
>>* second, do you think you will make and exe and service version for
>>Windows?
>>   =20
>>
>
>This is definitely something I would like to do, but because of other
>responsibilities, I won't have time to get to this for at least a few
>months. =20
>
> =20
>
>>The first in mainly for my knowledge, but the second is very important =
for
>>me. For now, I must ask my users - and you know they aren't all so clea=
ver -
>>not to close that black windows, and please don't ask why...
>>   =20
>>
>
>Another thing you can do is create a simple script that runs in the
>middle of the night when the users are not around.  For example, you
>can use the "Scheduled Tasks" feature in Windows to do this.
>
>For example, having the script call a few commands like
>
>auto_check
>send_message
>poll_passives
>process_message
>
>and then exit should be sufficient to have dibs do its thing.
>
> =20
>
>>I tried myself with py2exe, but as I'm not so cleaver either, I did not
>>succeed.=20
>>   =20
>>
>
>I tried that quickly a while ago and had some trouble as well.  I
>believe the problem is with the win32 extensions that are not a
>built-in part of python.
>
> =20
>
>>As I'm far from being a python guru, I will appreciated you help.
>>And, please, please, do not tell me that you give up this project! HiSp=
read
>>sucks!!
>>   =20
>>
>
>I have certainly not given up on the project, but the last couple of
>months (and probably the next month) have been very busy for me.  I
>hope to have more time to focus on DIBS again soon.  For the short
>term, however, trying the "scheduled tasks" idea that calls a
>script is probably the best approach.  Let me know if you have
>problems/questions about that or if that works for you.
>
>Also, you may want to post your questions to the mailing list or the
>forums at http://sourceforge.net/projects/dibs/.  Of course, I will
>still respond if you just want to email me personally, but the forums
>and mailing lists are read by other people who might have suggestions
>that I don't think of.
>
>Thank you for your interest in DIBS,
>-Emin=20
>
>
> =20
>


--=20
BrokenClock
-
Mon blog/My blog: http://brokenclock.free.fr
-
Warning:
Except explicit opposition of your side, any mail sent to me should be us=
ed on my blog, partly or entirely, including your name, but without your =
mail. If your opposition takes place after the post of a story on my blog=
, it will be deleted on request by mail.
--
Avertissement:
Sauf opposition explicite de votre part, tout mail qui me sera adress=E9 =
pourra =EAtre utilis=E9 dans mon blog, en partie ou en totalit=E9, y comp=
ris votre nom et =E0 l'exception de votre mail. Si votre opposition se ma=
nifestait apr=E8s la publication d'un article sur mon blog, celui-ci sera=
it retir=E9 sur simple demande par mail.

Re: [Dibs-discussion] DIBS Contract wanted for testing

[Emin M. <em...@al...>]

The best way to search for contracts is to either look at 

  http://martinian.com:8000/~emin/cgi-bin/peer_finder/show_contracts.cgi

for existing contracts or post your own contract using the
post_contract command to post your own contract server.  If you do
that and nobody else responds in a couple of days, I would be happy to
take it.

If the contract server annoys you for some reason or is difficult to
use, please let me know so that I can improve it.

Thanks,
-Emin

"J. Wren Hunt" <wr...@hu...> writes:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: RIPEMD160
>
> Hi!
>
> I am new to DIBS after just discovering it and was wondering if anyone
> would be interested in helping me establish a contract or two for
> testing purposes? If so I'd love to hear from you. I guess for starters
> I'm looking to play with a CD-sized chunk so about 700MB sounds like a
> good figure, although I don't anticipate transferring that much data all
> at once - just rather small files intermittently to avoid wasting bandwidth.
>
> My DIBS GPG key is included below. Thanks in advance for any and all help!
>
>
> - --
>
> Cheers!
>
> J. Wren Hunt
> Cambridge, MA. USA
>
> - ------------
> "In theory, there is no difference between theory and practice. But, in
> practice, there is." - Jan L.A. van de Snepscheut
>
> +------------------------------------------------------------------+
> | v-card   http://wrenhunt.homelinux.org/data/wren.vcf             |
> | x.509    http://wrenhunt.homelinux.org/data/thawte_wren_hunt.cer |
> | OpenPGP  ADF5 1432 A59E 8F4D 4AE7  4DFE 03FA 91E1 4A24 D6F4      |
> +------------------------------------------------------------------+
>
>
> Public key for 0xC317F1824BB35F56
> - -----BEGIN PGP PUBLIC KEY BLOCK-----
> Version: GnuPG v1.4.1rc2 (Darwin)
>
> mQGiBEIdI8ARBACH6C5yCr+ifa7pnI2fVpNd5NVukHbEVRq808o74vR9Xj/pyNkm
> VlWnZY3BEn2YDN+w6acghv88OMdscIcDsTXAtUo6Y8JvM8f0jAlRncZiDrZgBxoy
> dIyJVYGpDmaHAhzxMBvRgdxSWmQte2wsyob0LX9pZxVCAsJAbf9rQuLqLwCgy3hH
> 9/bu/ovGwhW9ddN4QQRtol0D/RQuypwRUPLM4eNC+Imr1jgvqQAiHaX5aZzhN/+X
> ILuvHZ0S0XxUDAGGl0ZDFGdbKBiK6BKu5YE/TatMuhVv6cclXyOjCKvVlMVYxZZ0
> s8Uml9xBVzdOXmN4MregCNsSm04zvbfjHSvPf66+loUw90aa7v1Ud/Ld5NByTwl6
> O6FDA/9ClvnTVVXmKibtlyzLbJwlMNNcMwEXeuIxCAOuFpQ6EF+JSHnVLDfRFufx
> 5KsNE1Thw0by+nJ8F+3b5vUVnNFbgsZ/yJmBHsOapxw0dsq0dB47dtoZM6kEG0iJ
> GHazvu81y8hsAlgtcYedGTV2z8wik3xbd256En37HPN6Y+opQbRDSi4gV3JlbiBI
> dW50IChmb3IgRElCUyB1c2Ugb25seSkgPHdyZW4uZGlic0B3cmVuaHVudC5ob21l
> bGludXgub3JnPohkBBMRAgAkBQJCHSPAAhsDBQkB4TOABgsJCAcDAgMVAgMDFgIB
> Ah4BAheAAAoJEMMX8YJLs19Wd6kAoLbzJBFQKvl8K32l9/IiB/42lU+TAJ0Xf5Eb
> g0r0Y4s8ZuXMrcyAvDTcu7kCDQRCHSPXEAgAoz7zqcvmFPVDXCZAZCNhTqXK/urn
> wnMESmetDEfbOEh72y3wS87VPqPugrrp5Y/P7nbUpp4sS0bF2MQVv2h83+PDK29x
> uRwiSqX6evrRIbKw7KDJP8vcMNFvB1ajta/eRtCs3sZGOnlcxmJ34EJcQAu2IsMP
> 7vM5twCychiN5MaJ3VP4ryt9l4BXnF6vjLj3wobbESI3vTOuv2vamZuPQQQQ0vpd
> 4lB9IPyNVPTgUyoVud3VxPzdLnIvfwq8P3QpeQN7qfJMYRUA0k34pMKUcTc9U2wm
> nvwNRjf8EPH7r+HAtCDVjD1Rs7S60tWqHLPwUt89W2le+jOtPn+Ae3V9RwADBQf/
> aaZt+nZrqygSjw9+bdUCuro+BecVa5LSmCNbFWFzLmeGKzQqHt7hl33sddHzsmhf
> iXF9/BZ/pDumFBJ7w1dUzHbMQRzIuVYHTs+xWZ6K/XteBeuRell+zJCWeXDfbNcS
> tcToDjvwfHxyAB0ZPRNpA0v3xWboW1q4SMkrpex0WCuCMxkMffCZoQmRwpszvI1I
> DSy2hIvB/XwT7AxpvaAzW/hHcC6RI9pWxtLlzGZBQQT7wCXGYAC7APAShjmNJRnt
> 8SsNuN+239OaoOE0O6QJLr9IyTn/FH2uWQ48ixVp9aitoEppNdnjJG1xnUb69HQf
> r6AEOOyJ0eETdLMPrGZnbohPBBgRAgAPBQJCHSPXAhsMBQkB4TOAAAoJEMMX8YJL
> s19WA8wAnifCrOKMOxAAWnzl5xyYWhw5TNTFAKCEpduKuBUItVaBvlYyEbv7wayI
> Tg==
> =8z3T
> - -----END PGP PUBLIC KEY BLOCK-----
>
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.1rc2 (Darwin)
>
> iD8DBQFCODVFA/qR4Uok1vQRA8VzAJ9x8F1lTsa0QLZ/FEXChasAk+V2lgCgnpU1
> dfHCxvFVx3x3OmT1eha2KBw=
> =5Cd7
> -----END PGP SIGNATURE-----
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> Dibs-discussion mailing list
> Dib...@li...
> https://lists.sourceforge.net/lists/listinfo/dibs-discussion

[Dibs-discussion] DIBS Contract wanted for testing

[J. W. H. <wr...@hu...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: RIPEMD160

Hi!

I am new to DIBS after just discovering it and was wondering if anyone
would be interested in helping me establish a contract or two for
testing purposes? If so I'd love to hear from you. I guess for starters
I'm looking to play with a CD-sized chunk so about 700MB sounds like a
good figure, although I don't anticipate transferring that much data all
at once - just rather small files intermittently to avoid wasting bandwidth.

My DIBS GPG key is included below. Thanks in advance for any and all help!


- --

Cheers!

J. Wren Hunt
Cambridge, MA. USA

- ------------
"In theory, there is no difference between theory and practice. But, in
practice, there is." - Jan L.A. van de Snepscheut

+------------------------------------------------------------------+
| v-card   http://wrenhunt.homelinux.org/data/wren.vcf             |
| x.509    http://wrenhunt.homelinux.org/data/thawte_wren_hunt.cer |
| OpenPGP  ADF5 1432 A59E 8F4D 4AE7  4DFE 03FA 91E1 4A24 D6F4      |
+------------------------------------------------------------------+


Public key for 0xC317F1824BB35F56
- -----BEGIN PGP PUBLIC KEY BLOCK-----
Version: GnuPG v1.4.1rc2 (Darwin)
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=8z3T
- -----END PGP PUBLIC KEY BLOCK-----

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1rc2 (Darwin)

iD8DBQFCODVFA/qR4Uok1vQRA8VzAJ9x8F1lTsa0QLZ/FEXChasAk+V2lgCgnpU1
dfHCxvFVx3x3OmT1eha2KBw=
=5Cd7
-----END PGP SIGNATURE-----

Re: [Dibs-discussion] user experience installing

[Emin M. <em...@al...>]

"Eric S. Johansson" <es...@in...> writes:

> by almost silently, I mean that it indicated that it did not import
> only by saying that count of keys imported was 0.  There was no other
> warning that the key import process failed.  So if I had not been
> paying attention I would have thought that the key had been imported
> one in reality it hadn't been.

I suspect that the key was already imported so the key import didn't
fail, it just didn't change anything.

> good point.  And by the way, I apologize for missed speech recognition
> errors.  Sometimes inverts the meaning of my sentences or otherwise
> alters them in ways designed to maximize mischief or confusion.

No problem.

> according to the instructions you give one could come away with the
> impression that if you have multiple machines, the e-mail address you
> use for key generation should be all the same address, in my case
> esj...@ha....  When in reality, you do not need even to have
> valid addresses in the keys, just unique ones hence my schema
> described earlier.

OK, I'll mention this point in the next set of docs.

> can you not specify different key rings?  It seems to me it would be
> basic functionality but again... assuming you can specify key rings,
> it should be relatively simple to create dibs specific key rings to
> prevent collision with user generated keyrings

I don't know how to do this.  I think the GPG philosophy is that keys
should be unique.

> I think with key naming, I should be all set.  I think that assuming
> that the asymmetrical key "e-mail address" really isn't an e-mail
> address but just an identifier with an @ in the middle works best.

Yes.

-Emin

Re: [Dibs-discussion] more errors

[Emin M. <em...@al...>]

"Eric S. Johansson" <es...@in...> writes:

> Processed msg in file
> /root/.dibs/incoming/1098377962201.0+y0JMSGLwFumoi4DXI2tzA.
> closing log
> Connection from ('192.168.0.53', 1057) closed.
> Doing periodic checks...
> Spawning auto_check.
>  opening log /root/.dibs/logfile at level 0
> WARNING: Unable to read file or directory
> "/home/esj/camram_core.working/modules/es...@re....32434:1081526411"
This looks like a stale NFS handle or some other weird file that isn't
really a file.  If that's the case, I don't know what else DIBS could
do.

> in store attempt.
> File or directory not stored.
>  Could not store file
> '/home/esj/postfix/postfix-2.2-20040801/src/anvil/anvil' using name
> '/autoBackup/home/esj/postfix/postfix-2.2-20040801/src/anvil/anvil'
>  due to exception: <dibs_lib.dibs_exceptions.NoPeersToTradeWith
> instance at 0x9c03c6c>
> Subject: <DIBS> DIBS ERROR
>  Command was '['/usr/bin/dibs.py', 'auto_check']'.
>  Exception of type '<class dibs_lib.dibs_exceptions.NoPeersToTradeWith
> at 0x94e70cc>':  'DIBSException: No peers exist in database with
> sufficient trading room.Traceback:

Well, you are out of storage space on your peers...

> and then it goes to hell from there leaving the passive side at 100%
> cpu and the last error:
>
> Subject: <DIBS> DIBS ERROR
>  Command was '['/usr/bin/dibs.py', 'poll_passives']'.
>  Exception of type '<class dibs_lib.dibs_exceptions.DIBSException at
> 0x9ee148c>':  'DIBSException: Unexpected socket error: (-2, Name or
> service not known)Traceback:
>   File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_daemon.py",
> line 322, in __init__
>     self.conn.connect((host,port))
> '.
>  Traceback:
>   File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_ui.py", line
> 645, in ?
>     c.run(sys.argv)
>   File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_ui.py", line
> 279, in run
>     c = dibs_daemon.DIBSClient(peer[0],peer[1].host,peer[1].port)
>   File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_daemon.py",
> line 336, in __init__
>     raise DIBSException, ('Unexpected socket error: (' +
>  mailing error msg to es...@ha...

This error "Name or service not known" is probably happening because
DIBS is trying to connect to a machine with a host name that doesn't
exist. 

> blech... any chance of specing storage as M and G for megabytes and
> gigabytes??  

Coming in the next release.

> I don't seem able to type enough zeros.  Alternative, any
> chance of running backups to within 20 percent of the specified
> partition?  

That's a good idea.  Submit a request for enhancement on sourceforge
and I'll put it on the list of things to do.

> Oh yes!  How about being able to specify the storage path
> for backing everything up?  (Don't recall seeing that in the
> documentation).

Not sure what you mean here.  You can use whatever links and path you
like in the autoBackup directory.  For example, make a link in
~/.dibs/autoBackup called foo/bar/baz to some/special/place and things
in some/special/place will be backed up as /autoBackup/foo/bar/baz/place.

> If I'm being too hard, I apologize.

Not at all, I appreciate the feedback.  The best way for DIBS to
improve is for people to use it and complain about the problems.  If
anything, I feel bad about releasing software which has such
problems.  At a company, I would never release software like that.
But as a single developer working on stuff during my spare time, I
find the feedback essential.

Thanks,
-Emin

[Dibs-discussion] [Fwd: <DIBS> DIBS ERROR]

[Eric S. J. <es...@in...>]

why is this considered an error?  To me it would be a warning you push
into syslog but basically ignored.  Yes something is wrong but it's
probably a symptom of something else such as insufficient storage (as
noted in earlier message)

Thinking about the storage problem, it's probably not uncommon to need
to add or delete storage so maybe being able to add storage as a Delta
might be a good feature from a usability perspective.

---eric


-------- Original Message --------
Subject: <DIBS> DIBS ERROR
Date: Thu, 21 Oct 2004 18:26:02 -0400 (EDT)
From: es...@ha...
To: undisclosed-recipients:;

Command was '['/usr/bin/dibs.py', 'process_message']'.

Exception of type '<class
dibs_lib.dibs_exceptions.DuplicateStorageRequestFromPeer at 0x8d2ac94>':
  'DIBSException: failed because it is already stored.
Traceback:
'.


Traceback:
   File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_ui.py", line
645, in ?
     c.run(sys.argv)
   File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_ui.py", line
476, in run
     d.ProcessMessage(file)
   File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_main.py", line
333, in ProcessMessage
     eval(cmdString)
   File "<string>", line 0, in ?
   File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_main.py", line
41, in StoreFileForPeer
     self.database.RememberStoringPieceForPeerOnOurMachine(dibsM,peer)
   File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_database.py",
line 476, in RememberStoringPieceForPeerOnOurMachine
     raise DuplicateStorageRequestFromPeer, dibsMsg



Offending file moved to /tmp/@22410.0.

[Dibs-discussion] more errors

[Eric S. J. <es...@in...>]

Processed msg in file 
/root/.dibs/incoming/1098377962201.0+y0JMSGLwFumoi4DXI2tzA.
closing log
Connection from ('192.168.0.53', 1057) closed.
Doing periodic checks...
Spawning auto_check.
 
opening log /root/.dibs/logfile at level 0
WARNING: Unable to read file or directory 
"/home/esj/camram_core.working/modules/es...@re....32434:1081526411" 
in store attempt.
File or directory not stored.
 
Could not store file 
'/home/esj/postfix/postfix-2.2-20040801/src/anvil/anvil' using name 
'/autoBackup/home/esj/postfix/postfix-2.2-20040801/src/anvil/anvil'
 due to exception: <dibs_lib.dibs_exceptions.NoPeersToTradeWith instance 
at 0x9c03c6c>
Subject: <DIBS> DIBS ERROR
 
Command was '['/usr/bin/dibs.py', 'auto_check']'.
 
Exception of type '<class dibs_lib.dibs_exceptions.NoPeersToTradeWith at 
0x94e70cc>':  'DIBSException: No peers exist in database with sufficient 
trading room.Traceback:
'.
 

and then it goes to hell from there leaving the passive side at 100% cpu 
and the last error:

Subject: <DIBS> DIBS ERROR
 
Command was '['/usr/bin/dibs.py', 'poll_passives']'.
 
Exception of type '<class dibs_lib.dibs_exceptions.DIBSException at 
0x9ee148c>':  'DIBSException: Unexpected socket error: (-2, Name or 
service not known)Traceback:
  File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_daemon.py", line 
322, in __init__
    self.conn.connect((host,port))
'.
 
 
Traceback:
  File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_ui.py", line 645, 
in ?
    c.run(sys.argv)
  File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_ui.py", line 279, 
in run
    c = dibs_daemon.DIBSClient(peer[0],peer[1].host,peer[1].port)
  File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_daemon.py", line 
336, in __init__
    raise DIBSException, ('Unexpected socket error: (' +
 
mailing error msg to es...@ha...

blech... any chance of specing storage as M and G for megabytes and 
gigabytes??  I don't seem able to type enough zeros.  Alternative, any 
chance of running backups to within 20 percent of the specified 
partition?  Oh yes!  How about being able to specify the storage path 
for backing everything up?  (Don't recall seeing that in the documentation).

If I'm being too hard, I apologize.  I drive myself even harder.  I've 
been this hard on myself for camram and IPCop 1.2.

---eric

Re: [Dibs-discussion] user experience installing

[Eric S. J. <es...@in...>]

Emin Martinian wrote:

>"Eric S. Johansson" <es...@in...> writes:
>
>  
>
>>yes but if you have the same e-mail address on all keys, it almost
>>silently refuses to import.  
>>    
>>
>
>almost?  So it works, but you are concerned something may be going
>wrong without GPG telling you?  
>  
>

by almost silently, I mean that it indicated that it did not import only 
by saying that count of keys imported was 0.  There was no other warning 
that the key import process failed.  So if I had not been paying 
attention I would have thought that the key had been imported one in 
reality it hadn't been.

>>Additionally, I really think they should be easier encapsulation or
>>better instructions on key management.  
>>    
>>
>
>Hmm, I'm not sure if this is a DIBS issue or a GPG issue.
>  
>

good point.  And by the way, I apologize for missed speech recognition 
errors.  Sometimes inverts the meaning of my sentences or otherwise 
alters them in ways designed to maximize mischief or confusion.

>  
>
>>For example, I'm creating all
>>my keys under the root account because that's how I run dibs.
>>however, if I have multiple requirements for root based asymmetrical
>>key systems, will gpg create multiple key pairs or will it just
>>overwrite?  My unfounded suspicion is overwrite. 
>>    
>>
>
>I'm not sure what you mean.  You can certainly have multiple keys with
>different names.  Do you want different keys with the same name?  That
>probably won't work, but why would you want that?
>  
>

according to the instructions you give one could come away with the 
impression that if you have multiple machines, the e-mail address you 
use for key generation should be all the same address, in my case 
esj...@ha....  When in reality, you do not need even to have 
valid addresses in the keys, just unique ones hence my schema described 
earlier.

>  
>
>>Therefore I would suggest it might be wise to store the dibs key
>>pair in ~/.dibs to keep it safe from overwriting etc.  I do realize
>>however that this may be easier said than done given that the gpg
>>realm is not exactly concerned with usability.
>>    
>>
>
>This seems like it could cause problems.  GPG wants any key you use to
>be in its keyring.
>  
>
can you not specify different key rings?  It seems to me it would be 
basic functionality but again... assuming you can specify key rings, it 
should be relatively simple to create dibs specific key rings to prevent 
collision with user generated keyrings

>It seems like I'm not understanding what you want to do with your
>keys.  If you want multiple machines to have the same private GPG key
>that is fine just give it the same name and import it everywhere.  If
>you want them to have different keys, give them different names.  Let
>me know if there is some other mode of operation you have in mind
>  
>
I think with key naming, I should be all set.  I think that assuming 
that the asymmetrical key "e-mail address" really isn't an e-mail 
address but just an identifier with an @ in the middle works best.

>>I think I figured out how to reproduce this failure.  On the
>>recipient machine, just give it a bad e-mail address for where to
>>e-mail messages.
>>    
>>
>
>Great!  Could you submit this as a bug report to sourceforge?  It's
>easier for me to track things that way.
>  
>

done

>This error means that machine A tried to store a file on machine B.
>Then later machine A decided it wanted to probe/remove/recover that
>file from machine B.  This error means that machine B never heard of
>the file machine A has in mind.  The most common way for this error to
>happen is that machine B gets killed (e.g., by Control-C or because
>machine A got turned off) while it was processing the message from A.
>
>Assuming you have enough redundancy in your backup you can basically
>ignore these messages.  Of course, if a lot of them happen often for
>no good reason it is a cause for concern.  In the future, I plan to
>have a feature where two peers can reconcile their database to figure
>out who was missing what when this error occurs, but that hasn't
>happened yet.
>
>Alternatively, you can issue the clear command and backup things from
>scratch.  That might be painful depending on how much you've already
>backed up.
>  
>

since I'm just running tests between two machines to make sure I can get 
things to work, it's easier for me to run clear command and just play 
around.

---eric

Re: [Dibs-discussion] user experience installing

[Emin M. <em...@al...>]

"Eric S. Johansson" <es...@in...> writes:

> yes but if you have the same e-mail address on all keys, it almost
> silently refuses to import.  

almost?  So it works, but you are concerned something may be going
wrong without GPG telling you?  

> What I'm slowly evolving to (circling the
> drain :-) is something like <host>.dibs@<domain> for the e-mail key
> for public keys on each machine but tell add-peer to use my e-mail
> address as where to e-mail data.

Yes, that would work.

> Additionally, I really think they should be easier encapsulation or
> better instructions on key management.  

Hmm, I'm not sure if this is a DIBS issue or a GPG issue.

> For example, I'm creating all
> my keys under the root account because that's how I run dibs.
> however, if I have multiple requirements for root based asymmetrical
> key systems, will gpg create multiple key pairs or will it just
> overwrite?  My unfounded suspicion is overwrite. 

I'm not sure what you mean.  You can certainly have multiple keys with
different names.  Do you want different keys with the same name?  That
probably won't work, but why would you want that?

> Therefore I would suggest it might be wise to store the dibs key
> pair in ~/.dibs to keep it safe from overwriting etc.  I do realize
> however that this may be easier said than done given that the gpg
> realm is not exactly concerned with usability.

This seems like it could cause problems.  GPG wants any key you use to
be in its keyring.

It seems like I'm not understanding what you want to do with your
keys.  If you want multiple machines to have the same private GPG key
that is fine just give it the same name and import it everywhere.  If
you want them to have different keys, give them different names.  Let
me know if there is some other mode of operation you have in mind.

> I was probably being too impatient.  I would have expected that the
> startup of the demon would have looked, seen that I had a link to
> /home and started copying.  This did not happen.  In fact, this did
> not happen multiple times.

Point taken.  I modified things in the upcoming version so that as
soon as the daemon is started it spawns an auto_check.

> I think I figured out how to reproduce this failure.  On the
> recipient machine, just give it a bad e-mail address for where to
> e-mail messages.

Great!  Could you submit this as a bug report to sourceforge?  It's
easier for me to track things that way.

> Command was '['/usr/bin/dibs.py', 'process_message']'.
> 
> Exception of type '<class
> dibs_lib.dibs_exceptions.RemoteFileNonExistent at 0xa16f7b4>':
> 'DIBSException: No record of storing piece '
> Version: GnuPG v1.2.4 (GNU/Linux)

This error means that machine A tried to store a file on machine B.
Then later machine A decided it wanted to probe/remove/recover that
file from machine B.  This error means that machine B never heard of
the file machine A has in mind.  The most common way for this error to
happen is that machine B gets killed (e.g., by Control-C or because
machine A got turned off) while it was processing the message from A.

Assuming you have enough redundancy in your backup you can basically
ignore these messages.  Of course, if a lot of them happen often for
no good reason it is a cause for concern.  In the future, I plan to
have a feature where two peers can reconcile their database to figure
out who was missing what when this error occurs, but that hasn't
happened yet.

Alternatively, you can issue the clear command and backup things from
scratch.  That might be painful depending on how much you've already
backed up.

-Emin

Re: [Dibs-discussion] consuming lots of CPU

[Eric S. J. <es...@in...>]

Emin Martinian wrote:

>"Eric S. Johansson" <es...@in...> writes:
>
>  
>
>>Set up initially to push green machine (trial) home directories out to
>>DMZ machine (redweb).  Setting up both machines as dibs daemons did
>>nothing.  They both sat there not doing anything.  
>>    
>>
>
>Did you have stuff in the auto backup directory?  The daemons should
>check the auto backup directory periodically (I think every 10
>minutes) depending on the value of the daemonTimeout or variable or
>something like that.
>  
>

I was probably being too impatient.  I would have expected that the 
startup of the demon would have looked, seen that I had a link to /home 
and started copying.  This did not happen.  In fact, this did not happen 
multiple times.

>>So I killed off the
>>demon on trial and ran  dibs.py auto_check.  This moved the files over
>>to redweb but unfortunately, after it is sensibly completed file
>>movement, it just sat there consuming 50 to 100 percent CPU.  
>>    
>>
>
>What do you mean by "it"? 
>
the process running dibs.py. auto_check.  I think I figured out how to 
reproduce this failure.  On the recipient machine, just give it a bad 
e-mail address for where to e-mail messages.  After tries to send the 
first message, and hangs.  After I fixed the e-mail address I started 
getting error messages like:

Command was '['/usr/bin/dibs.py', 'process_message']'.

Exception of type '<class dibs_lib.dibs_exceptions.RemoteFileNonExistent at 0xa16f7b4>':  'DIBSException: No record of storing piece '
Version: GnuPG v1.2.4 (GNU/Linux)
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=
=EZlg
'.
Traceback:
'.


Traceback:
  File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_ui.py", line 645, in ?
    c.run(sys.argv)
  File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_ui.py", line 476, in run
    d.ProcessMessage(file)
  File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_main.py", line 333, in ProcessMessage
    eval(cmdString)
  File "<string>", line 0, in ?
  File "/usr/lib/python2.2/site-packages/dibs_lib/dibs_main.py", line 75, in UnstoreFileForPeer
    raise RemoteFileNonExistent, pieceName



Offending file moved to /tmp/@3619.0.

Re: [Dibs-discussion] user experience installing

[Eric S. J. <es...@in...>]

Emin Martinian wrote:

>>for example, if you decide to use the same e-mail name for all keys on
>>all machines, exporting the user's public key from the public key ring
>>before importing anything is probably a good idea.
>>    
>>
>
>Does GPG complain otherwise?  I thought GPG is OK with importing keys
>which already exist.
>  
>

yes but if you have the same e-mail address on all keys, it almost 
silently refuses to import.  What I'm slowly evolving to (circling the 
drain :-) is something like <host>.dibs@<domain> for the e-mail key for 
public keys on each machine but tell add-peer to use my e-mail address 
as where to e-mail data.

Additionally, I really think they should be easier encapsulation or 
better instructions on key management.  For example, I'm creating all my 
keys under the root account because that's how I run dibs.  however, if 
I have multiple requirements for root based asymmetrical key systems, 
will gpg create multiple key pairs or will it just overwrite?  My 
unfounded suspicion is overwrite.  Therefore I would suggest it might be 
wise to store the dibs key pair in ~/.dibs to keep it safe from 
overwriting etc.  I do realize however that this may be easier said than 
done given that the gpg realm is not exactly concerned with usability.

---eric

Re: [Dibs-discussion] consuming lots of CPU

[Emin M. <em...@al...>]

"Eric S. Johansson" <es...@in...> writes:

> Set up initially to push green machine (trial) home directories out to
> DMZ machine (redweb).  Setting up both machines as dibs daemons did
> nothing.  They both sat there not doing anything.  

Did you have stuff in the auto backup directory?  The daemons should
check the auto backup directory periodically (I think every 10
minutes) depending on the value of the daemonTimeout or variable or
something like that.

> So I killed off the
> demon on trial and ran  dibs.py auto_check.  This moved the files over
> to redweb but unfortunately, after it is sensibly completed file
> movement, it just sat there consuming 50 to 100 percent CPU.  

What do you mean by "it"?  The sending machine moved files from it to
the receiver and then the sender seemed stalled or the receiver seemed
stalled?  One possibility is that the receiver was busy decoding what
the sender had sent and therefore the sender the was blocking until it
could talk to the receiver.  

By the way, you might want to set logLevel=-50 in your dibsrc.py file
to get more detailed log messages.

> and after about 5 minutes of letting things run on a relatively empty
> home directory, it still sitting at 90+% CPU utilization.
>
> obviously, I have messed something up.  Any ideas?

One thing you could do is to look in the sub-directories of
~/.dibs/outgoing/ and use the du -sk command to see if they are
growing.  If they are, this tells you DIBS is creating store messages
to send but just hasn't gotten around to sending them yet.

-Emin

Re: [Dibs-discussion] user experience installing

[Emin M. <em...@al...>]

"Eric S. Johansson" <es...@in...> writes:

> here are some notes from my installation experiences
>
> Red Hat fedora 1 and Gentoo
>
> setup.py: gets real cranky if you don't have documentation tools you
> count on installed.  

OK, I'll make the version of setup.py released in the next install be
more tolerant of missing doc tools.

> unfortunately dibsPublicKey dibsPrivateKey description really doesn't
> make sense.  How does a e-mail address translate to a location of the
> keys?  

You shouldn't have to specify a location.  If you just specify the
email address for the key, GPG should be able to get it out of your
key ring.

> if you mean it's the e-mail address keyed to the key pair, then say so.

OK, I'll put in a documentation bug so I remember to be clearer about this.

> Also, if you own all of the machines that you're backing up to/from, all
> of the keys will have the same e-mail address.  Not good.  

Why not?

> for example, if you decide to use the same e-mail name for all keys on
> all machines, exporting the user's public key from the public key ring
> before importing anything is probably a good idea.

Does GPG complain otherwise?  I thought GPG is OK with importing keys
which already exist.

> more as I trip over them.

OK. 

Thanks,
-Emin

Re: [Dibs-discussion] consuming lots of CPU

[Eric S. J. <es...@in...>]

Christian Stork wrote:

>Hi Eric,
>
>On Wed, Oct 20, 2004 at 04:19:49PM -0400, Eric S. Johansson wrote:
>... 
>  
>
>>and after about 5 minutes of letting things run on a relatively empty 
>>home directory, it still sitting at 90+% CPU utilization.
>>
>>obviously, I have messed something up.  Any ideas?
>>    
>>
>
>Unfortunately, I don't think that you messed something up.  We tried to
>perform small backups with DIBS a while back and had similar
>experiences.  If you feel like profiling the daemon and finding out
>where exactly it's spending its time (I didn't get to that yet) we could
>provide Emin with a hint where the main problem lies.
>  
>

ugh.  30 seconds of looking reveals I need to stick the last code 
fragment of dibs_ui.py into a function and call that function from 
dibs.py.  after I take a run...

couple thoughts: has anybody run ethereal watching this condition and 
could it be some sort of a unreleased lock condition.  The reason behind 
my two questions is that I have seen multithreaded code burn lots of 
cycles especially when there is a busy wait loop waiting on a resource 
lock.  Also, it's entirely possible that the two nodes are just talking 
back and forth to each as a result of some protocol failure and never 
ever terminate the protocol.  This condition would be detected most 
easily by looking at the traffic with ethereal.

just a couple thoughts.  No guarantee that they are even vaguely correct.

---eric

Re: [Dibs-discussion] consuming lots of CPU

[Christian S. <dib...@cs...>]

Hi Eric,

On Wed, Oct 20, 2004 at 04:19:49PM -0400, Eric S. Johansson wrote:
... 
> and after about 5 minutes of letting things run on a relatively empty 
> home directory, it still sitting at 90+% CPU utilization.
> 
> obviously, I have messed something up.  Any ideas?

Unfortunately, I don't think that you messed something up.  We tried to
perform small backups with DIBS a while back and had similar
experiences.  If you feel like profiling the daemon and finding out
where exactly it's spending its time (I didn't get to that yet) we could
provide Emin with a hint where the main problem lies.

-- 
Chris Stork   <>  Support eff.org!  <>   http://www.ics.uci.edu/~cstork/
OpenPGP fingerprint:  B08B 602C C806 C492 D069  021E 41F3 8C8D 50F9 CA2F

[Dibs-discussion] consuming lots of CPU

[Eric S. J. <es...@in...>]

obviously something has run away.

My test machines are on my own network.  One is on a DMZ and is passive 
listener and talker and the other is on the green network and can 
connect to the DMZ machine at will.

Set up initially to push green machine (trial) home directories out to 
DMZ machine (redweb).  Setting up both machines as dibs daemons did 
nothing.  They both sat there not doing anything.  So I killed off the 
demon on trial and ran  dibs.py auto_check.  This moved the files over 
to redweb but unfortunately, after it is sensibly completed file 
movement, it just sat there consuming 50 to 100 percent CPU.  hitting ^C 
gives me the following trace back:


Could not store file 
'/home/scm/.mozilla/default/pqksxpdt.slt/Cache/_CACHE_003_' using name 
'/autoBackup/home/scm/.mozilla/default/pqksxpdt.slt/Cache/_CACHE_003_'
 due to exception: <exceptions.KeyboardInterrupt instance at 0x4056490c>
Subject: <DIBS> DIBS ERROR

Command was '['/usr/bin/dibs.py', 'auto_check']'.

Exception of type '<class exceptions.KeyboardInterrupt at 0x402a238c>':  ''.


Traceback:
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_ui.py", line 645, 
in ?
    c.run(sys.argv)
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_ui.py", line 296, 
in run
    file))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
503, in StoreAsNecessary
    self.StoreFile(file,storageName)
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
594, in StoreFile
    raise e

Subject: <DIBS> DIBS ERROR

Command was '['/usr/bin/dibs.py', 'auto_check']'.

Exception of type '<class exceptions.KeyboardInterrupt at 0x402a238c>':  ''.


Traceback:
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_ui.py", line 645, 
in ?
    c.run(sys.argv)
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_ui.py", line 296, 
in run
    file))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
491, in StoreAsNecessary
    item)))
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
503, in StoreAsNecessary
    self.StoreFile(file,storageName)
  File "/usr/lib/python2.3/site-packages/dibs_lib/dibs_main.py", line 
594, in StoreFile
    raise e

mailing error msg to tri...@ha...
Could not send message from tri...@ha... to 
tri...@ha... because of exception:

<socket.error instance at 0x40568d4c>
Not saving database.
closing log

and after about 5 minutes of letting things run on a relatively empty 
home directory, it still sitting at 90+% CPU utilization.

obviously, I have messed something up.  Any ideas?

---eric

sorry for the double post: Re: [Dibs-discussion] user experience installing

[Eric S. J. <es...@in...>]

thought I was subscribed under one address and not the other.

--- eric

[Dibs-discussion] user experience installing

[Eric S. J. <es...@in...>]

here are some notes from my installation experiences

Red Hat fedora 1 and Gentoo

setup.py: gets real cranky if you don't have documentation tools you
count on installed.  The easy solution was to comment out all
documentation related elements within setup.py.  Try to find the RPMs
for the missing pieces was just too much work.  Gentoo made it possible
to find missing pieces after some searching.

building gpg key: built into default location (~root/.gnupg/...) with
e-mail address esj...@ha...

unfortunately dibsPublicKey dibsPrivateKey description really doesn't
make sense.  How does a e-mail address translate to a location of the
keys?  if you mean it's the e-mail address keyed to the key pair, then
say so.

Also, if you own all of the machines that you're backing up to/from, all
of the keys will have the same e-mail address.  Not good.  Should
probably change the documentation to handle the case of one entity
owning all machines in the backup cluster.

for example, if you decide to use the same e-mail name for all keys on
all machines, exporting the user's public key from the public key ring
before importing anything is probably a good idea.

more as I trip over them.

---eric