[ckan-dev] Help! - data resources gone missing (Aaron McGlinchy)

Fri Oct 24 06:52:30 UTC 2014

Hi all,

I'm also running against this bug in CKAN 2.2b, which is very disconcerting
for my end users as they lose attached files without a chance of getting
them back.
Has a bug already been reported?

I managed to work around the bug in a very specific case - I
programmatically upload working versions of annual reports to our CKAN, and
as the upload iterates over the 16 attached resources, some of them
sometimes get marked as "deleted". As I have configured the upload, I've
got the resource_id - my end users don't have this luxury! My workaround is
to retrieve all resource dicts (gettable api, resource_show), set their
status to "active", retrieve the parent dataset's dict (package_show), then
update the parent's dictionary's "resources" key with the list of resource
dictionaries, and finally update the package using this modified package
dictionary. A short write-up is here:
http://data-demo.dpaw.wa.gov.au/dataset/report-automation/resource/ad551de0-5ef2-49e9-8b33-62de737a5aca

Cheers,
Florian

Dept Parks & Wildlife, Western Australia

On Tue, Jun 10, 2014 at 1:16 PM, Nigel Babu <nigel.babu at okfn.org> wrote:

> Hi Aaron,
>
> Yes, it does sound like it's a bug. Can you file one against
> github.com/ckan/ckan? I'll try to reproduce the issue locally meanwhile
> and
> confirm what's going on.
>
> On Mon, Jun 09, 2014 at 03:39:21AM +0000, Aaron McGlinchy wrote:
> > Bump on my query as to whether this is a further bug (after we applied
> the fix for the first issue it seemed to be resolved, but similar issue
> arose when deleting resources).
> >
> > ------------------------------
> >
> > Message: 3
> > Date: Wed, 28 May 2014 22:42:36 +0000
> > From: Aaron McGlinchy <McGlinchyA at landcareresearch.co.nz>
> > To: "ckan-dev at lists.okfn.org" <ckan-dev at lists.okfn.org>
> > Subject: Re: [ckan-dev] Help! - data resources gone missing
> > Message-ID:
> >         <
> A924326C14714F409FB3703B815FF9FD1225F672 at HERMES.landcare.ad.landcareresearch.co.nz
> >
> >
> > Content-Type: text/plain; charset="us-ascii"
> >
> > We updated the datapusher to this commit
> d433bd8a20ffd82ea684dc0c2a63bfc11dda5ada.
> >
> > This seemed to fix the problem of resources disappearing during
> uploading of other resources.  However today I just found that when I went
> into a dataset that had multiple (6) resources, and decided to delete one
> of them (the 3rd in the list in this case), on completion of the delete
> resources 4,5 and 6 had disappeared as well.
> >
> > When I deleted the resource, I went into Manage | Edit resource, clicked
> on the Remove button, and then clicked the Delete button.  Possibly
> removing the file/link first then deleting has triggered an additional
> error.
> >
> > Does this sound like a further bug?
> >
> > Cheers
> > Aaron
> > ------------------------------
> >
> > Message: 2
> > Date: Tue, 20 May 2014 23:39:58 +0000
> > From: Aaron McGlinchy <McGlinchyA at landcareresearch.co.nz>
> > To: "ckan-dev at lists.okfn.org" <ckan-dev at lists.okfn.org>
> > Subject: Re: [ckan-dev] Help! - data resources gone missing
> > Message-ID:
> >         <
> A924326C14714F409FB3703B815FF9FD1225C186 at HERMES.landcare.ad.landcareresearch.co.nz
> >
> >
> > Content-Type: text/plain; charset="us-ascii"
> >
> > Aaargh!  It has happened again, this time I did not attempt to reorder
> resources (I did edit some resource descriptions, and spontaneousy all but
> one of the approx 5 resources I had uploaded disappeared).
> >
> > Nigel - can you please point me to the link for the branch you think
> fixes this issue, I will ask my Tech guy to implement it (is it difficult
> to do/any instructions I can point him too)?
> >
> > Thanks
> > Aaron
> >
> > ------------------------------
> >
> > Message: 2
> > Date: Mon, 19 May 2014 09:44:42 +0530
> > From: Nigel Babu <nigel.babu at okfn.org>
> > To: CKAN Development Discussions <ckan-dev at lists.okfn.org>
> > Subject: Re: [ckan-dev] Help! - data resources gone missing
> > Message-ID:
> >         <CALGFUD7NG8UDo1+cBeBbTGv=
> chmYXTuMsottzdD370pdYoxGBg at mail.gmail.com>
> > Content-Type: text/plain; charset="utf-8"
> >
> > Hi Alex and Aaron,
> >
> > I believe the latest commit in the stable branch fixes this issue. I'll
> arrange for an email out to ckan-announce today.
> >
> > Nigel Babu
> > Developer, Open Knowledge
> >
> >
> > On 19 May 2014 09:10, Aaron McGlinchy <McGlinchyA at landcareresearch.co.nz
> >wrote:
> >
> > >  We are also on CKAN 2.2.
> > >
> > >
> > >
> > > Like Alex the files that disappeared have had their state set to
> > > deleted, however setting it back to active and then switching the
> > > dataset from private-public-private did not make them appear again.
> > > The files that disappeared on me we some large zip files, a couple of
> jpgs, tsv, and pdf ?
> > > not limieted to csv files as Alex experienced.
> > >
> > >
> > >
> > > I noticed that one of the files position integer setting was
> > > duplicated, so I changed it (the xlsx file) to 12 to avoid a conflict,
> > > but the resources still did not show up (apart from 1).  Have decided
> > > to delete the original dataset and reload, being careful to put in the
> > > order I want so I don?t need to try reordering which seems to trigger
> the problem.
> > >
> > >
> > >
> > > [image: cid:image002.png at 01CF7368.09A36550]
> > >
> > >
> > >
> > > ------------------------------
> > >
> > >
> > >
> > > Message: 3
> > >
> > > Date: Sat, 17 May 2014 17:52:25 +1000
> > >
> > > From: "Alex (Maxious) Sadleir" <maxious at gmail.com>
> > >
> > > To: CKAN Development Discussions <ckan-dev at lists.okfn.org>
> > >
> > > Subject: Re: [ckan-dev] Help! - data resources gone missing
> > >
> > > Message-ID:
> > >
> > >                 <
> > > CAL3EF=9P=JbdfZJ4DL3aVKnYWC_uZqvY=Qsy3Kqi+oJ84ROA-g at mail.gmail.com>
> > >
> > > Content-Type: text/plain; charset=UTF-8
> > >
> > >
> > >
> > > Haven't upgraded datapusher since the 2.2 release so fbc6dc3 on stable
> > > branch.
> > >
> > >
> > >
> > > I see there are some fixes so will deploy and see if it is resolved,
> > > maybe if this is a known severe issue let ckan-announce know?
> > >
> > >
> > >
> > > On Thu, May 15, 2014 at 3:38 PM, Nigel Babu <nigel.babu at okfn.org>
> wrote:
> > >
> > > > This sounds terrible. Can you tell me what's the latest commit in
> > > > the
> > >
> > > > datapusher version you're using? I remember hearing about this bug
> > > > and
> > >
> > > > us fixing it.
> > >
> > > >
> > >
> > > > Nigel Babu
> > >
> > > > Developer, Open Knowledge
> > >
> > > >
> > >
> > > >
> > >
> > > > On 15 May 2014 08:23, Alex (Maxious) Sadleir <maxious at gmail.com>
> wrote:
> > >
> > > >>
> > >
> > > >> I have experienced this too - my best theory is it happens to
> > >
> > > >> datasets with very small CSV files and there is some kind of race
> > >
> > > >> condition in the DataPusher updating the resources.
> > >
> > > >>
> > >
> > > >> On datasets that it occurs (7 or so small CSVs), it occurs
> > >
> > > >> consistently after it occurs for the first time whenever another
> > > >> file
> > >
> > > >> is added or existing file edited. Interestingly the resources are
> > > >> all
> > >
> > > >> marked as deleted within ms of each other and if I manually set the
> > >
> > > >> state back to Active in postgres, I have to set the dataset Private
> > >
> > > >> and Public to trigger them to reappear.
> > >
> > > >>
> > >
> > > >> On Thu, May 15, 2014 at 12:43 PM, Aaron McGlinchy
> > >
> > > >> <McGlinchyA at landcareresearch.co.nz> wrote:
> > >
> > > >> > Hi,
> > >
> > > >> >     I created a dataset and had uploaded a number of resources
> > >
> > > >> > yesterday (about 10, of various file formats).
> > >
> > > >> >
> > >
> > > >> > Today I had some more resources to add, and I edited the dataset
> > >
> > > >> > and began uploading the additional resources.  Whilst doing so I
> > >
> > > >> > used the reorder resources button to arrange them in the order I
> > >
> > > >> > wanted.  When I did this with the first additional resource it
> > >
> > > >> > worked fine, but when I added another resource and went to
> > > >> > reorder,
> > >
> > > >> > the list of resources to rearrange that came up only had a single
> > >
> > > >> > resource (the one just loaded) and all the ones I had loaded
> > > >> > earlier
> > > disappeared.
> > >
> > > >> >
> > >
> > > >> > When I clicked back to the Dataset it only showed that single
> > > resource.
> > >
> > > >> > However I knew the link for some of the resources I had
> > > >> > previously
> > >
> > > >> > loaded so I opened that and the resource I check is still there,
> > >
> > > >> > just not showing up under the dataset.
> > >
> > > >> >
> > >
> > > >> > What can I do to get them back listed under the dataset?
> > >
> > > >> >
> > >
> > > >> > Thanks
> > >
> > > >> > Aaron
> >
> > ________________________________
> >
> > Please consider the environment before printing this email
> > Warning: This electronic message together with any attachments is
> confidential. If you receive it in error: (i) you must not read, use,
> disclose, copy or retain it; (ii) please contact the sender immediately by
> reply email and then delete the emails.
> > The views expressed in this email may not be those of Landcare Research
> New Zealand Limited. http://www.landcareresearch.co.nz
> > _______________________________________________
> > ckan-dev mailing list
> > ckan-dev at lists.okfn.org
> > https://lists.okfn.org/mailman/listinfo/ckan-dev
> > Unsubscribe: https://lists.okfn.org/mailman/options/ckan-dev
>
> --
> Nigel Babu
> Developer, Open Knowledge
> _______________________________________________
> ckan-dev mailing list
> ckan-dev at lists.okfn.org
> https://lists.okfn.org/mailman/listinfo/ckan-dev
> Unsubscribe: https://lists.okfn.org/mailman/options/ckan-dev
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/ckan-dev/attachments/20141024/9a799cd0/attachment-0002.html>