Time
S
Nick
Message
05:23 djbrooke joined #dataverse
09:21 djbrooke joined #dataverse
12:36 donsizemore joined #dataverse
12:38 donsizemore
@pdurbin re https://github.com/IQSS/dataverse-ansible/issues/20 i don't get the joeboo-ansible not found error, i bumped dataverse-ansible to 4.9.2 and tweaked the jdbc driver step
12:39 donsizemore
@pdurbin though i'm unable to publish a dataset. looking into why now.
12:44 pdurbin
donsizemore: thanks. I used to share an office with sjoeboo, by the way. Nice guy. Small world. :)
12:44 donsizemore
@pdurbin found a permissions snafu with solr. re-running playbook
12:55 donsizemore
@pdurbin oh, it was the prefix/protocol/separator dealy in 4.9.
13:25 pdurbin
yeah, that thing
13:25 pdurbin
donsizemore: you're fixing it?
13:27 donsizemore
@pdurbin i'm stumbling towards fixing it. the permissions thing cropped up with solr 7, the shoulder/authority stuff rearranged with 4.9, and admittedly my summer has been crazy-pants
13:33 donsizemore
so, i can publish a dataset now, and sure enough, i can't upload files (jhove errors)
13:33 pdurbin
wow, so you're able to replicate the bug Jamie found
13:34 donsizemore
i still have those two jhove file copies in there from 4.8... whatever. i bet i don't need them any more (the ansible run extracts from the current dvinstall.zip)
13:34 pdurbin
the jhove files should definitely be in that zip
13:36 donsizemore
the role just needed some love. again with the Twilight Zone Stopwatch...
13:40 pdurbin
:)
13:43 donsizemore
okay, this time i'm using whatever jhove.conf came in dvinstall.zip: /editFilesFragment.xhtml @200,65 fileUploadListener="#{EditDatafilesPage.handleFileUpload}": java.lang.NoClassDefFoundError: edu/harvard/hul/ois/jhove/viewer/ConfigWindow javax.el.ELException: /editFilesFragment.xhtml @200,65 fileUploadListener="#{EditDatafilesPage.handleFileUpload}": java.lang.NoClassDefFoundError: edu/harvard/hul/ois/jhove/viewer/ConfigWindow
13:44 donsizemore
but without the role placing jhove.conf, there isn't one in the instance deployed in vagrant
13:48 pdurbin
donsizemore: huh. Does https://github.com/IQSS/dataverse/commit/ce38642 help at all?
13:50 donsizemore
they're getting unzipped into /tmp but install nor setup-* reference them. i updated the jhove config task to place both current files; i'll nuke the old copy in the role.
13:51 pdurbin
donsizemore: do you use the Perl script (install) in Ansible?
13:51 donsizemore
dataverse-ansible doesn't call "./install" outright so there was some deconstruction
13:51 pdurbin
ah, you anticipated my question :)
13:51 pdurbin
`install` was heavily refactored in 4.9.2
13:51 donsizemore
it would've been simpler to populate the .config file and all the script, but i forget why i didn't/couldn't do that
13:52 donsizemore
^call the script
13:54 donsizemore
actually, that step had a directory path out of order, it never should've worked
13:59 pdurbin
I'm not sure what the .config file is.
14:13 donsizemore
the default.config(?) file the perl installer checked for a non-interactive run
14:14 pdurbin
oh oh, default.config. thanks
14:14 donsizemore
git-push-and-go-get-coffee! i think we're good. or at least better.
14:17 pdurbin
https://github.com/IQSS/dataverse-ansible/commit/69aa3db looks good to me
14:18 pdurbin
donsizemore: you even replied to Jamie on the GitHub issue. Thanks!
14:31 donsizemore
ah do sumthin' raht, wunst in a while
14:33 pdurbin
:)
14:34 pdurbin
wicked smaaht
15:16 donsizemore
@pdurbin i'm off to help move my mom. you have a great weekend!
15:25 pdurbin
you too
15:43 pameyer joined #dataverse
15:52 pameyer
@donsizemore - default.config didn't exist when you created that ansible role; might be one reason you didn't use it
15:59 pdurbin
pameyer: good point
16:00 pdurbin
pameyer: question about https://github.com/IQSS/dataverse/issues/4558
16:17 pameyer
pdurbin: shoot
16:19 pdurbin
Do you think it was caused by https://github.com/IQSS/dataverse/pull/4038 ?
16:21 pameyer
possibly, possibly not - initially had assumed it was; some digging made me suggest otherwise
16:21 pameyer
but not definative
16:22 pdurbin
Ok. I'm wondering if you or andrewSC have time to work on a pull request.
16:23 pameyer
thought that was an iqss dev one?
16:24 pdurbin
It didn't make it into the sprint. I tried.
16:24 pameyer
my memory is horrible
16:24 pameyer
it's not high on my list at the moment, but that may change depending on how/if date stuff interacts with other migration stuff
16:24 pdurbin
mine too. that's why iqlogbot is here :)
16:25 pdurbin
Cool. Maybe you'll look at date stuff and get bothered by how lots of dates aren't being indexed.
16:28 pameyer
I do practice annoyance-driven development sometimes
16:28 pdurbin
heh
16:32 pdurbin
pameyer: does /usr/local/glassfish4/glassfish/domains/domain1/files/10.5072/FK2/1OVPCW/FK2/1OVPCW make sense as an upload directory for DCM? I'm a little troubled by the two "FK2" folders in there.
16:32 pameyer
pdurbin: it doesn't; looks like that's a sign of old DCM stuff
16:32 pdurbin
Ok. I'm using the mock.
16:32 pdurbin
Should I fix the mock?
16:33 pameyer
ah - don't think mock was fixed
16:33 pameyer
heading out for food
16:33 * pdurbin
is hungry
18:03 pameyer
pdurbin: back from food
18:03 pameyer
do you think the mock is worth fixing?
18:08 pdurbin
pameyer: the jury is still out. I'm playing around with testDcmChecksumValidationMessages in DatasetsIT at the moment.
18:09 pdurbin
The package file isn't being created and I'm not sure why.
18:09 pdurbin
This used to work.
18:12 pameyer
"This used to work" - sounds like my yesterday
18:18 pdurbin
pameyer: this is what I'm getting: Dataset doi:10.5072/FK2/BLQUPZ is not locked for DCM upload. Exiting
18:18 pdurbin
Job Failed. See Log for more information.
18:19 pdurbin
Caught throwable in main execution loop with Throwable message: java.io .IOException: Dataset doi:10.5072/FK2/BLQUPZ is not locked for DCM upload, and stack trace: com.ibm.jbatch.container.exception.BatchContainerRuntimeException: java.io.IOException: Dataset doi:10.5072/FK2/BLQUPZ is not locked for DCM upload
18:19 pdurbin
Caused by: java.io .IOException: Dataset doi:10.5072/FK2/BLQUPZ is not locked for DCM upload
18:19 pdurbin
at edu.harvard.iq.dataverse.batch.jobs.importer.filesystem.FileRecordJobListener.beforeJob(FileRecordJobListener.java:177)
18:20 pdurbin
I'm on 58dbb9d of the 4610-upload-dual-mode branch.
18:21 pameyer
haven't tried that w\ a real DCM; but it shouldn't matter - then locks should be happening on the dv side
18:23 pdurbin
Hmm, I wonder if it's this: Dataset Identifier (datasetId=1645): FK2/BLQUPZ
18:35 pdurbin
pameyer: for now I disabled the check for that lock and I was able to create a package file
18:35 pdurbin
I'm not sure if we should be worried about this lock stuff or not.
18:37 pameyer
it wouldn't make much sense to send "checksums ok" if the dataset hadn't been locked for upload; but I'm not sure if that inverts failing if not
18:37 pdurbin
yeah, I dunno
18:38 pdurbin
not trying to worry you
18:38 pdurbin
you're paranoid enough :)
18:38 pdurbin
it's helpful for me to have a package file in my installation
18:38 pdurbin
as I work on the dual mode front end stuff
19:02 pdurbin
pameyer: I pushed some fixes for the tests. Next question. I'm hearing that DCM can reply with "I'm down" or "I'm busy". When does it reply with "I'm busy"? What is it busy doing? Down I understand. :)
19:06 pameyer
?
19:07 pameyer
it sounds like when I tried to explain things, I made them more confusing
19:07 pameyer
the "I'm down" signal is no response from the DCM
19:08 pameyer
there's no "I'm busy" signal; but the 404 response to sr.py can mean "try again in a few minutes because I'm working" or "I'm partly broken and need support intervention"
19:11 pdurbin
pameyer: ok. Thanks. I'll see what I can do.
20:14 pameyer joined #dataverse
20:23 pdurbin
heading out soon. have a good weekend, everybody
20:28 pdurbin left #dataverse
21:23 pameyer
pdurbin: for when you check iqlogbot https://github.com/sbgrid/data-capture-module/tree/wksp_mock201808/api
21:26 pameyer left #dataverse
21:59 jri joined #dataverse
23:00 jri joined #dataverse
