Time
S
Nick
Message
04:57 djbrooke joined #dataverse
06:41 Haridass joined #dataverse
06:41 Haridass
Hi Dataverse Team
06:42 Haridass
I am facing one issue unzipping glassfish in Ubuntu
06:42 Haridass
We are setting up Dataverse in our Ubuntu server as per the steps given in http://guides.dataverse.org/en/latest/developers/ubuntu.html
06:43 Haridass
I have successfully downloaded glassfish using the command
06:43 Haridass
•wget http://download.java.net/glassfish/4.1/release/glassfish-4.1.zip
06:43 Haridass
But when I try to unzip using
06:43 Haridass
unzip glassfish-4.1*zip
06:43 Haridass
I am getting the following error
06:44 Haridass
End-of-central-directory signature not found. Either this file is not a zipfile, or it constitutes one disk of a multi-part archive. In the latter case the central directory and zipfile comment will be found on the last disk(s) of this archive.
07:05 jri joined #dataverse
10:38 pdurbin
I just downloaded that glassfish zip file (`md5sum glassfish-4.1.zip` shows "2fd41ad9af8d41d1c721c1b25191f674") and it unzipped fine.
13:15 bsilverstein joined #dataverse
13:23 djbrooke joined #dataverse
13:24 djbrooke joined #dataverse
13:27 donsizemore joined #dataverse
13:28 donsizemore
@pdurbin: is https://spaces.internet2.edu/display/InCFederation/Error+Handling+Service worth creating a github issue (and want me to do so?)
13:30 pdurbin
donsizemore: huh. Maybe. First time I've seen that page though so I'm not sure what the GitHub issue would be about.
13:30 pdurbin
"The goal of Federated Error Handling is to provide a better user experience in those situations where an IdP provides an SP with insufficient information (attributes) to make an access control decision."
13:31 pdurbin
That sounds like me! The poor Service Provider (SP) who doesn't get enough information. :)
13:34 djbrooke joined #dataverse
13:35 donsizemore
@pdurbin: as is, Dataverse tells us exactly which attribute is missing, but a more user-friendly error URL could smooth things over down the road
14:05 pameyer joined #dataverse
14:10 djbrooke joined #dataverse
14:28 pdurbin
yeah, that's a good idea
14:34 Haridass joined #dataverse
14:35 Haridass
Hi dataverse team
14:35 pdurbin
Haridass: hi! Did you see my reply at http://irclog.iq.harvard.edu/dataverse/2016-06-14#i_36835 ?
14:37 Haridass
yes pdurbin
14:38 Haridass
glassfish got downloaded successfully
14:38 Haridass
that issue got resolved
14:39 pdurbin
oh good
14:39 Haridass
But I have another issue
14:39 Haridass
While installing Dataverse I am getting the following error
14:39 Haridass
nope I haven't been able to connect to the local instance oof postgresql as admin user
14:52 djbrooke joined #dataverse
14:53 djbrooke joined #dataverse
15:04 djbrooke joined #dataverse
15:13 axfelix joined #dataverse
15:15 pdurbin
Haridass: following error? Did you paste an error?
15:18 Haridass
the error is
15:19 Haridass
Nope, I haven't been able to connect to the local instance of postgresql as admin user
15:19 Haridass
If you can provide me your email, I can sen you the screen shot
15:19 pdurbin
You can email it to support
15:19 Haridass
ok
15:20 Haridass
By the way, can I know your name
15:20 Haridass
or shall I address it as Support Team
15:20 pdurbin
Philip Durbin but most people call me Phil. http://www.iq.harvard.edu/people/philip-durbin
15:21 Haridass
ok Phil
15:21 pdurbin
Haridass: you're installing Dataverse on a server, right? You aren't setting up a dev environment, are you?
15:32 Haridass
Yes you are correct
15:32 Haridass
As a first step I want to install Dataverse and use it as a end user.
15:33 Haridass
We are planning to build a team to set the dev environment for customization work in the future.
15:33 pdurbin
Make sense. If you want, you could try `vagrant up` from your Ubuntu box.
15:35 pdurbin
Haridass: great that you're building a team of people do development on Dataverse!
15:37 Haridass
is Vagranr Up something like a VM virtual box manager
15:37 pdurbin
Haridass: I see your screenshots in https://help.hmdc.harvard.edu/Ticket/Display.html?id=237133 . For now, since you're just playing around, you might want to set Postgres to "trust", if that makes sense.
15:37 axfelix
@Hardiass: yup, it's exactly that :)
15:37 Haridass
"trust" done already
15:38 Haridass
still same error
15:38 pdurbin
To use Vagrant, you install Vagrant and VirtualBox and then run `vagrant up` to spin up a CentOS VM in VirtualBox running Dataverse.
15:39 djbrooke joined #dataverse
15:39 pdurbin
Haridass: the error is telling you that postgres is not running.
15:39 Haridass
the link https://help.hmdc.harvard.edu/Ticket/Display.html?id=237133 is asking for login
15:39 Haridass
do I need to register for help.hmdc.harvard.edu
15:40 Haridass
"trust" done already, still same error
15:41 Haridass
Postgres is running and I am able run SQL statements in psql
15:41 pdurbin
huh
15:41 pdurbin
I wonder why the installer thinks postgres isn't running.
15:41 Haridass
there we 2 screenshots in my email
15:42 Haridass
the first one is where I have to specify certain values for dataverse installation
15:42 Haridass
are those values correct
15:43 Haridass
there are 2 screenshots in my email
15:44 Haridass
please verify the values are correct in my first screenshot
15:44 Haridass
postgres is having only one user 'postgres' with password 'excel'
15:45 Haridass
but the dataverse installer is asking for Postgres ADMIN Password as well as Postgres User Password
15:47 Haridass
Also
15:48 Haridass
my PC is a windows machine with Ubuntu in VM
15:48 Haridass
Ubuntu on virtual disk
15:48 Haridass
in this setup can I use vagrant up
15:49 djbrooke joined #dataverse
15:50 pdurbin
Oh, no. You would run `vagrant up` from Windows. You'd install Vagrant and VirtualBox on your Windows box.
15:56 Haridass
I will get Vargrant Up set up tommorow and will let you know the progress then.
15:56 Haridass
in the meantine will this chat history be maintained
15:57 pdurbin
yep
15:57 pdurbin
I back it up. MySQL dump
15:57 Haridass
do I need logins for https://help.hmdc.harvard.edu/Ticket/Display.html?id=237133
15:58 pameyer
pdurbin: mind if I move some of your stubs around in the dcm repo?
16:02 pdurbin
pameyer: please feel free!
16:02 pdurbin
BBQ time. Back in a while.
16:02 pameyer
enjoy :)
16:03 djbrooke joined #dataverse
16:26 majesticc joined #dataverse
17:01 djbrooke joined #dataverse
17:29 djbrooke joined #dataverse
17:40 djbrooke joined #dataverse
17:45 pdurbin
so much BBQ. and ice cream
17:48 pdurbin
Haridass: I just wrote back explaining how to log into the ticketing system.
17:58 djbrooke joined #dataverse
18:11 pdurbin
pameyer: I'm working on the actionlogrecord stuff. You're actively checking this table from time to time? There's no API or GUI to do so.
18:14 djbrooke joined #dataverse
18:18 djbrooke joined #dataverse
18:28 donsizemore joined #dataverse
18:29 pameyer
pdurbin: what's the context for the actionlogrecord question?
18:31 pdurbin
pameyer: on Thursday I added a todo to RequestRsyncScriptCommand saying "make sure the error is logged to the actionlogrecord". That's still a requirement, right?
18:32 pameyer
the error being logged somewhere is - but is actionlogrecord where operational errors for the admins would usually go?
18:34 pdurbin
Errors *can* be logged there. I'm not sure how often anyone looks at it, since (again) there's no GUI or API to look at the actionlogrecord table.
18:34 pdurbin
pameyer: we log lots of errors to server.log. You're familiar with server.log, right?
18:35 pameyer
vaguely
18:35 pdurbin
It's the main Glassfish log.
18:36 pameyer
is there an expected place for admins to look for information about things going wrong?
18:36 pameyer
it's sounding like that might not be actionlogrecord
18:37 pdurbin
I'd say server.log is the main place to look for errors. But maybe that's just me.
18:39 pameyer
think it's worth asking the ops/curation people on your end where they usually look?
18:39 pdurbin
let's ask donsizemore since he runs a production Dataverse installation
18:40 pdurbin
pameyer: curation people here don't have access to the database (actionlogrecord, etc.) and don't use APIs.
18:43 pdurbin
bricas: do you have opinions on the "actionlogrecord" table?
18:50 pdurbin
pameyer: I also have a todo that says "notify user on any failure" and I might need a little more detail on that.
18:50 donsizemore
@pdurbin i generally look in server.log or syslog - aside from turning on debugging as needed
18:50 pameyer
donsizemore: thanks
18:51 pdurbin
donsizemore: ok, and debugging often involves cranking up the verbosity in server.log
18:51 pameyer
pdurbin: that'll be failure for post-upload validation
18:51 donsizemore
@pdurbin in my one year of experience, Dataverse pretty much spits out the most helpful bits to the screen while Glassfish buries it in a straw pile of Java exceptions
18:51 pameyer
aka - dcm needs enough info to send a message to the user (and curators) that there was a problem
18:52 pdurbin
donsizemore: I guess that's good. We try to show useful errors in the UI. :)
18:54 pdurbin
pameyer: so the DCM will send a message to Dataverse about if the upload succeeded or failed. A JSON message, I assume. Have you defined what the JSON will look like yet?
18:54 pameyer
only to the level of "something that dataverse is happy getting messages in"
18:55 pdurbin
Dataverse is happy with JSON .
18:56 pameyer
is there an API for "tell user $x somthing"?
18:56 pameyer
I didn't find one in a quick look at the docs - but that problem could be on my end
18:57 djbrooke joined #dataverse
19:02 pdurbin
pameyer: it sounds like you want an API endpoing called "notifications" or something. There is no such thing.
19:02 pdurbin
endpoint*
19:03 pameyer
is that because there's another way of doing it?
19:03 pameyer
or is this something to put on the list of requirements?
19:05 pdurbin
pameyer: it looks like we have "listen for DCM messages" already at https://docs.google.com/spreadsheets/d/1gZy4w21xnJsKRvOtarl_GUb_aeS53Nk7kJ04NYfi10E/edit?usp=sharing
19:06 pameyer
yup
19:07 pdurbin
pameyer: that reminds me. I have a script for you to try to run. A project management script. :)
19:08 pameyer
I can probably make it generate *interesting* errors ;)
19:08 pameyer
where does it need to run?
19:08 pdurbin
pameyer: please try running this: https://gist.github.com/pdurbin/5a98932c94055674278ebf47d241782e
19:10 pameyer
doesn't *look* like a rootkit - hang on a few
19:10 pdurbin
heh
19:11 pameyer
looks to me like it worked - want me to send you the output?
19:12 pdurbin
that's ok
19:13 pdurbin
pameyer: but maybe some day we can use a similar script to export our task list into a document
19:14 majestic joined #dataverse
19:15 pameyer
cool
19:15 pdurbin
that one was getting complicated enough that a script became helpful
19:18 pameyer
might be helpful other places as well
19:18 pameyer
and I learned that google will give you csv
19:19 pdurbin
:)
20:03 bjonnh
pdurbin: http://guides.dataverse.org/en/latest/installation/prep.html#hardware-requirements
20:03 bjonnh
I found that to give me an idea of what the harvard instance is
20:04 bjonnh
is there a way to see a stats page for the harvard instance, to know how much they store right now?
20:05 pdurbin
bjonnh: no, we don't have a stats page but if you email support
20:06 pdurbin
and I actually have to run, sorry
20:06 bjonnh
ohh that's all?
20:06 pdurbin left #dataverse
20:06 bjonnh
see you :)
20:07 pameyer
bjonnh - how big were you hoping for?
20:07 bjonnh
I have two projects in mind
20:08 bjonnh
one university wide, one NIH national-center wide
20:08 bjonnh
but I'm sure we're not going to reach 6TB before a couple of years
20:09 bjonnh
because our field is lagging way behind others in term of datasharing
20:10 pameyer
mind if I ask which field?
20:34 bjonnh
pharmacognosy
20:34 bjonnh
pharmacy/natural products
20:47 metamattj joined #dataverse
20:53 djbrooke joined #dataverse
20:54 djbrooke_ joined #dataverse
20:57 djbrooke joined #dataverse
20:58 djbrooke_ joined #dataverse
21:30 pameyer
bjonnh: cool - not a field I've got enough overlap with to give you informed guesses as to data volume though
21:31 pameyer
but you're probably in a better spot to do that than I am anyhow
21:36 djbrooke joined #dataverse
21:38 bjonnh
well most publication/grant related datasets are not that huge
21:39 bjonnh
except that we have cell-imaging and huge metabolomic mass spectrometry data files that are coming with their tens to hundreds of GB
21:39 bjonnh
but these are marginal for now
21:41 pameyer
that gets into the range where http uploads might cause problems
21:42 pameyer
part of the reason we're working on non-http uploads
