Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-25 Thread Tomeu Vizoso
On Tue, Aug 25, 2009 at 17:42, Martin Langhoffmartin.langh...@gmail.com wrote:
 On Thu, Aug 20, 2009 at 6:56 PM, Martin
 Langhoffmartin.langh...@gmail.com wrote:
 On olpc-sur we're seeing reports of my journal contents disappeared
 after reboot. (Thread starts at
 http://lists.laptop.org/pipermail/olpc-sur/2009-August/004177.html )

 So, we have a couple of users who could send a sample datastore to a
 developer. I think Tomeu is interested in diagnosing and fixing the
 problem within Sugar (great!) -- if so, do you want me to ask for a
 zipfile of the corrupt datastore on your behalf?

If this is easier than getting logs, then yes.

 It would also be very important for these Uy users to get a rescue my
 lost journal activity. Anyone interested in helping?

I will be happy to answer any questions about this.

Regards,

Tomeu

 cheers,



 m
 --
  martin.langh...@gmail.com
  mar...@laptop.org -- School Server Architect
  - ask interesting questions
  - don't get distracted with shiny stuff  - working code first
  - http://wiki.laptop.org/go/User:Martinlanghoff
 ___
 Sugar-devel mailing list
 sugar-de...@lists.sugarlabs.org
 http://lists.sugarlabs.org/listinfo/sugar-devel




-- 
«Sugar Labs is anyone who participates in improving and using Sugar.
What Sugar Labs does is determined by the participants.» - David
Farning
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-21 Thread Tomeu Vizoso
On Fri, Aug 21, 2009 at 07:46, Martin Langhoffmartin.langh...@gmail.com wrote:
 On Fri, Aug 21, 2009 at 12:24 AM, Andrés Ambroisandresambr...@gmail.com 
 wrote:
 I'll help in any way I can. Would an upload webservice similar to Sacha's for
 uploading the necessary datastore bits help? We can then distribute a simple
 cli tool that uploads the data.

 Thanks! If Uy was using School Servers, this wouldn't be an issue. The
 DS data would be there :-/

 For end users, I don't think a network-based solution is best. A local
 activity will be a lot more effective.

 For debugging we might need to get some sample 'broken' datastores
 (with due care for user privacy, etc...).

 Will collecting the indexes be enough or do you need the whole datastore?

 For debugging, better to get the whole thing I guess.

Yes, though if we manage to get the logs dir of a machine just after
the DS was lost, there's lots of chances we are going to be able to
determine the exact cause for the loss. Then we would be in a better
position to find the best fix.

Regards,

Tomeu

-- 
«Sugar Labs is anyone who participates in improving and using Sugar.
What Sugar Labs does is determined by the participants.» - David
Farning
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-21 Thread Martin Langhoff
2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
 Yes, though if we manage to get the logs dir of a machine just after
 the DS was lost, there's lots of chances we are going to be able to

The ones on the ramdisk? :-(



m
-- 
 martin.langh...@gmail.com
 mar...@laptop.org -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-21 Thread Tomeu Vizoso
2009/8/21 Martin Langhoff martin.langh...@gmail.com:
 2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
 Yes, though if we manage to get the logs dir of a machine just after
 the DS was lost, there's lots of chances we are going to be able to

 The ones on the ramdisk? :-(

Is ~/.sugar/default/logs in the ramdisk? This would be the first time
I hear about it.

If it's like that, maybe Andres could modify a XO to have that dir in
the nand and wait for it to happen?

Regards,

Tomeu

 m
 --
  martin.langh...@gmail.com
  mar...@laptop.org -- School Server Architect
  - ask interesting questions
  - don't get distracted with shiny stuff  - working code first
  - http://wiki.laptop.org/go/User:Martinlanghoff




-- 
«Sugar Labs is anyone who participates in improving and using Sugar.
What Sugar Labs does is determined by the participants.» - David
Farning
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-21 Thread Andrés Ambrois
On Friday 21 August 2009 05:08:28 am Martin Langhoff wrote:
 2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
  Yes, though if we manage to get the logs dir of a machine just after
  the DS was lost, there's lots of chances we are going to be able to

 The ones on the ramdisk? :-(

  I think ~/.sugar/default/logs

  We're holding a jam the 29th, so maybe someone affected could attend, 
otherwise it'll be very hard for teachers to upload large files (ADSL broadband 
has a 128 Kbps uplink here in .uy). 

  I'll post a mail in Sur and olpc-uruguay, after I get some sleep. 


 m

-- 
  -Andrés
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-21 Thread Andrés Ambrois
On Friday 21 August 2009 05:10:21 am Tomeu Vizoso wrote:
 2009/8/21 Martin Langhoff martin.langh...@gmail.com:
  2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
  Yes, though if we manage to get the logs dir of a machine just after
  the DS was lost, there's lots of chances we are going to be able to
 
  The ones on the ramdisk? :-(

 Is ~/.sugar/default/logs in the ramdisk? This would be the first time
 I hear about it.

 If it's like that, maybe Andres could modify a XO to have that dir in
 the nand and wait for it to happen?

Remember that Ceibal XOs have root access locked-down. And I recently found 
out that since the key-delegation stuff was implemented, we can't request 
developer keys. Not from OLPC at least, and LATU is not providing that service 
that I know...

I'll also try to flash an XO with 8.2.1 and try to reproduce. 

 Regards,

 Tomeu

  m
  --
   martin.langh...@gmail.com
   mar...@laptop.org -- School Server Architect
   - ask interesting questions
   - don't get distracted with shiny stuff  - working code first
   - http://wiki.laptop.org/go/User:Martinlanghoff

-- 
  -Andrés
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-21 Thread Martin Langhoff
2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
 2009/8/21 Martin Langhoff martin.langh...@gmail.com:
 2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
 Yes, though if we manage to get the logs dir of a machine just after
 the DS was lost, there's lots of chances we are going to be able to

 The ones on the ramdisk? :-(

 Is ~/.sugar/default/logs in the ramdisk? This would be the first time
 I hear about it.

OLPC's builds (at least the F7/F9 based series) I understood they were
on a ramdisk. Looking at it now, however, it seems that they aren't --
but a restart of X clears it out anyway.

So by the time the problem is _seen_, *poof*, logs are gone.

Maybe something in the sugar startup clears it out. Dunno. Doesn't
seem to be olpc-configure.



m
-- 
 martin.langh...@gmail.com
 mar...@laptop.org -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-21 Thread Martin Dengler
On Fri, Aug 21, 2009 at 10:36:44AM +0200, Martin Langhoff wrote:
 2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
  2009/8/21 Martin Langhoff martin.langh...@gmail.com:
  2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
  Yes, though if we manage to get the logs dir of a machine just after
  the DS was lost, there's lots of chances we are going to be able to
 
  The ones on the ramdisk? :-(
 
  Is ~/.sugar/default/logs in the ramdisk? This would be the first time
  I hear about it.
 
 OLPC's builds (at least the F7/F9 based series) I understood they were
 on a ramdisk. Looking at it now, however, it seems that they aren't --
 but a restart of X clears it out anyway.
 
 So by the time the problem is _seen_, *poof*, logs are gone.

You sure that ~/.sugar/default/logs doesn't just get renamed to
~/.sugar/default/logs/N?  That's what happens now, but perhaps not
on older builds?

 m

Martin


pgpPvhO3kuQzs.pgp
Description: PGP signature
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-21 Thread Tomeu Vizoso
2009/8/21 Martin Langhoff martin.langh...@gmail.com:
 2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
 2009/8/21 Martin Langhoff martin.langh...@gmail.com:
 2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
 Yes, though if we manage to get the logs dir of a machine just after
 the DS was lost, there's lots of chances we are going to be able to

 The ones on the ramdisk? :-(

 Is ~/.sugar/default/logs in the ramdisk? This would be the first time
 I hear about it.

 OLPC's builds (at least the F7/F9 based series) I understood they were
 on a ramdisk. Looking at it now, however, it seems that they aren't --
 but a restart of X clears it out anyway.

 So by the time the problem is _seen_, *poof*, logs are gone.

 Maybe something in the sugar startup clears it out. Dunno. Doesn't
 seem to be olpc-configure.

Should be kept in a subdir in there, from:

http://git.sugarlabs.org/projects/sugar/repos/mainline/blobs/cbd0509f4bcfe1b3e52f8898b1cb83d522337ab8/src/logsmanager.py#line25

But anyway, the failure to read the DS would be in the first boot
where the journal appears empty, not in the boot where the DS index
got corrupted.

Regards,

Tomeu



 m
 --
  martin.langh...@gmail.com
  mar...@laptop.org -- School Server Architect
  - ask interesting questions
  - don't get distracted with shiny stuff  - working code first
  - http://wiki.laptop.org/go/User:Martinlanghoff




-- 
«Sugar Labs is anyone who participates in improving and using Sugar.
What Sugar Labs does is determined by the participants.» - David
Farning
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-21 Thread Martin Langhoff
2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
 Should be kept in a subdir in there, from:

Damn! I wish I had known about this! So 3 old dirs are kept - excellent.

 But anyway, the failure to read the DS would be in the first boot
 where the journal appears empty, not in the boot where the DS index
 got corrupted.

Well, if you can get your hands on the 'corrupt' datastore dir,
doesn't it also contain the xapian DB? Just renaming it into
'datastore' should re-produce the error...

Maybe not?



m
-- 
 martin.langh...@gmail.com
 mar...@laptop.org -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-21 Thread Tomeu Vizoso
2009/8/21 Martin Langhoff martin.langh...@gmail.com:
 2009/8/21 Tomeu Vizoso to...@sugarlabs.org:
 Should be kept in a subdir in there, from:

 Damn! I wish I had known about this! So 3 old dirs are kept - excellent.

 But anyway, the failure to read the DS would be in the first boot
 where the journal appears empty, not in the boot where the DS index
 got corrupted.

 Well, if you can get your hands on the 'corrupt' datastore dir,
 doesn't it also contain the xapian DB? Just renaming it into
 'datastore' should re-produce the error...

 Maybe not?

I think it should if we know the exact build image on which to place
it. So I would say whatever is easier for people to provide.

Regards,

Tomeu



 m
 --
  martin.langh...@gmail.com
  mar...@laptop.org -- School Server Architect
  - ask interesting questions
  - don't get distracted with shiny stuff  - working code first
  - http://wiki.laptop.org/go/User:Martinlanghoff




-- 
«Sugar Labs is anyone who participates in improving and using Sugar.
What Sugar Labs does is determined by the participants.» - David
Farning
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-20 Thread Tomeu Vizoso
On Thu, Aug 20, 2009 at 18:56, Martin Langhoffmartin.langh...@gmail.com wrote:
 On olpc-sur we're seeing reports of my journal contents disappeared
 after reboot. (Thread starts at
 http://lists.laptop.org/pipermail/olpc-sur/2009-August/004177.html )

 This reminds me a lot of the issues we saw earlier in development of
 what became 8.2.x -- if for any reason Sugar doesn't like the Journal
 contents or indexes, it moves the 'datastore' directory aside, and
 starts anew.

 That's my hunch, and I am hoping to hear confirmation from the users
 reporting the problem.

 Two pleas for help

  - A sugar activity that does some data-recovery and freeing up the space.

 The good news is that we have a json file with the metadata for each
 Journal entry, so we can either load it back into the Journal (mstone
 had published a cli script that exports/imports to the Journal?) or
 create Journal Entry Bundles and save them to a USB stick.

 I suspect the JEB path might help users more -- if the Journal is
 choking on some specific metadata, a script that reloads it all in one
 go will make it choke again. Having files as JEBs allows step-by-step
 attempts.

How do you envision the UI of the tool that will recover the JEBs?

There's a good chance that the xapian index got corrupted because the
machine lost power while the index on disk was on an inconsistent
state. If that's the case, just rebuilding the whole index should work
fine.

Would be nice to have more details about the failure so we can both
fix it and propose the best tool for the job. Maybe a ceibal jam
volunteer with some linux knowledge would like to work with us on
this?

Regards,

Tomeu

  - Diagnosis.

 If it's something we can pin down, it'll be great to have a fix! This
 will surely involve getting in touch with teachers on olpc-sur.

 cheers,



 m
 --
  martin.langh...@gmail.com
  mar...@laptop.org -- School Server Architect
  - ask interesting questions
  - don't get distracted with shiny stuff  - working code first
  - http://wiki.laptop.org/go/User:Martinlanghoff
 ___
 Sugar-devel mailing list
 sugar-de...@lists.sugarlabs.org
 http://lists.sugarlabs.org/listinfo/sugar-devel




-- 
«Sugar Labs is anyone who participates in improving and using Sugar.
What Sugar Labs does is determined by the participants.» - David
Farning
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-20 Thread Martin Langhoff
On Thu, Aug 20, 2009 at 7:44 PM, Tomeu Vizosoto...@sugarlabs.org wrote:
 How do you envision the UI of the tool that will recover the JEBs?

We don't need many words there :-) all we need is

 - A listing of Journals available (by date - the numbers after
'datastore' are epoch seconds IIRC) - if there's only one, skip this
stage...

 - A view that looks a bit like the Journal view... we could show an
'action' that is 'inject into journal' (icon of the activity, arrow,
journal icon?).

 There's a good chance that the xapian index got corrupted because the
 machine lost power while the index on disk was on an inconsistent
 state. If that's the case, just rebuilding the whole index should work
 fine.

In that case, another option is to not have a UI, and just merge all
the datastores present, in one go, no more questions asked. As part of
it, files are moved (instead of cp'd).

Not sure how to make the above atomic  safe.

 Would be nice to have more details about the failure so we can both
 fix it and propose the best tool for the job. Maybe a ceibal jam
 volunteer with some linux knowledge would like to work with us on
 this?

Many are reading olpc-sur -- I hope someone might help.

cheers,



m
-- 
 martin.langh...@gmail.com
 mar...@laptop.org -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-20 Thread Andrés Ambrois
On Thursday 20 August 2009 06:58:48 pm Martin Langhoff wrote:
 On Thu, Aug 20, 2009 at 7:44 PM, Tomeu Vizosoto...@sugarlabs.org wrote:
  Would be nice to have more details about the failure so we can both
  fix it and propose the best tool for the job. Maybe a ceibal jam
  volunteer with some linux knowledge would like to work with us on
  this?

 Many are reading olpc-sur -- I hope someone might help.

I'll help in any way I can. Would an upload webservice similar to Sacha's for 
uploading the necessary datastore bits help? We can then distribute a simple 
cli tool that uploads the data.

Will collecting the indexes be enough or do you need the whole datastore?

 cheers,



 m

-- 
  -Andrés
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-20 Thread James Cameron
On Thu, Aug 20, 2009 at 06:56:17PM +0200, Martin Langhoff wrote:
 On olpc-sur we're seeing reports of my journal contents disappeared
 after reboot. (Thread starts at
 http://lists.laptop.org/pipermail/olpc-sur/2009-August/004177.html )

Hmm.  I've two XOs with 8.2.1 being used by children of a friend, and
one of them reported a loss event recently ... what do I look for in the
filesystem to check for evidence of this?

-- 
James Cameron
http://quozl.linux.org.au/
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel


Re: [Sugar-devel] RFH - Journal corruption reports fom 8.2.1 users in Uy

2009-08-20 Thread Martin Langhoff
On Fri, Aug 21, 2009 at 12:24 AM, Andrés Ambroisandresambr...@gmail.com wrote:
 I'll help in any way I can. Would an upload webservice similar to Sacha's for
 uploading the necessary datastore bits help? We can then distribute a simple
 cli tool that uploads the data.

Thanks! If Uy was using School Servers, this wouldn't be an issue. The
DS data would be there :-/

For end users, I don't think a network-based solution is best. A local
activity will be a lot more effective.

For debugging we might need to get some sample 'broken' datastores
(with due care for user privacy, etc...).

 Will collecting the indexes be enough or do you need the whole datastore?

For debugging, better to get the whole thing I guess.


m
-- 
 martin.langh...@gmail.com
 mar...@laptop.org -- School Server Architect
 - ask interesting questions
 - don't get distracted with shiny stuff  - working code first
 - http://wiki.laptop.org/go/User:Martinlanghoff
___
Devel mailing list
Devel@lists.laptop.org
http://lists.laptop.org/listinfo/devel