Hi Hans,
Thanks for the advice! Sadly, I think our situations might not be the same.
We've been using external auth for a long time, so it's definitely not a new
change, and the user in question created their account after this was in place.
We log our users in with just their IU username, and tack the "@indiana.edu"
onto it with Apache to create the email address after authenticating the
username through IU. I checked the galaxy_user table, and there doesn't seem to
be any accidental duplicates or funny business with the user name.
User is active, external, not deleted, not purged, and we don't enforce any
disk quotas.
I pulled the histories associated with the user id:
select * from history where user_id = '95' ;
id | create_time | update_time | user_id |
name | hid_counter | deleted | purged | genome_build | importable |
slug | published | importing
------+----------------------------+----------------------------+---------+--------------------+-------------+---------+--------+--------------+------------+------+-----------+-----------
1665 | 2014-02-23 18:42:23.497483 | 2014-03-11 19:04:55.614609 | 95 |
Unnamed history | 5 | t | f | dm3 | f
| | f | f
1722 | 2014-02-28 20:46:50.292015 | 2014-03-11 19:06:47.498152 | 95 |
RNAseq | 7 | f | f | ? | f
| | f | f
1811 | 2014-03-11 19:04:55.910381 | 2014-03-11 19:13:06.82213 | 95 |
Unnamed history | 2 | f | f | ? | f
| | f | f
(3 rows)
I took a look at "RNAseq":
select * from history_dataset_association where history_id = '1722';
id: 10027
history_id: 1722
dataset_id: 7258
create_time: 2014-02-28 20:51:32.369667
update_time: 2014-03-06 15:37:27.033951
copied_from_history_dataset_association_id: 9778
hid: 4
name :
https://usegalaxy.org/datasets/bbd44e69cb8906b52e29eb7a7c6c4900/display?to_ext=
peek:@HWI-B5-690_0086_FC:2:1:2711:936#CNNNNN/1 ... (this checks out with the
datafile)
extension: txt
metadata: {"data_lines": 145087556, "dbkey": "dm3"}
deleted: f
visible: t
purged: f
extended_metadata_id: uploaded txt file
I found this file in the dataset_files and there doesn't appear to be anything
wrong about it. It appears that the datasets in question were downloaded from
Galaxy main (by the looks of the upload URL).
I'm at a loss. I'll ask the user to clear all their browser information just in
case this is something simple like that. I'll ask them to re-upload those
files. In the meantime, any other sleuthing suggestions would be greatly
appreciated.
Sincerely,
Carrie Ganote
________________________________________
From: Hans-Rudolf Hotz [h...@fmi.ch]
Sent: Wednesday, March 12, 2014 5:07 AM
To: Ganote, Carrie L
Cc: galaxy-dev@lists.bx.psu.edu
Subject: Re: [galaxy-dev] History seems to load forever
Hi Carrie
Was the history which is causing troubles created before you switched to
external authentication? Or has the user a slightly different e-mail
address now? eg different use of capital letters, like 'foo.bar@home'
and 'Foo.Bar@home' - we had such problems, when we switched to external
authentication.
I had to dig into our MySQL database and fix the "galaxy_user" table. In
your case, I recommend looking at the "galaxy_user" table and check
whether the user has more than one entry. Take the "id" (ie the number)
and search the "history" table using the id as 'user_id'. I would guess
you can fix your problem by changing the 'user_id' in the "history"
table for the history which is causing troubles.
BUT, BE VERY CAREFUL! you can do a lot of damage to the database. Switch
off the galaxy server first, and MAKE a back up the database before you
change anything.
Regards, Hans-Rudolf
On 03/11/2014 11:58 PM, Ganote, Carrie L wrote:
> Hi List,
>
> I just pulled and merged from galaxy-dist (I think it's dc067a95261d),
> updated my database, and migrated my tools. The issues I am able to see
> are minor, but I have a user who can no longer access a particular
> history. It apparently just hangs. If they delete the history, Galaxy
> responds, but it will hang again when they try to add back any datasets.
> Unfortunately, I don't see any errors in the paster logs about this. I
> fear that their dataset entries in the Postgresql db have been corrupted
> during the update, or perhaps the files themselves were purged. I cannot
> reproduce their issues with my own histories - I'm able to bring up
> saved histories and upload etc. I don't know if they can share a history
> if it can't be loaded. How should I troubleshoot this?
> I'm set up to authenticate through apache, which makes it harder to get
> in as the user and see what's up. Is there an easy way to find out what
> histories belong to what users? I can search the database?
>
> Any advice would be greatly appreciated!
>
> -Carrie
>
>
> ___________________________________________________________
> Please keep all replies on the list by using "reply all"
> in your mail client. To manage your subscriptions to this
> and other Galaxy lists, please use the interface at:
> http://lists.bx.psu.edu/
>
> To search Galaxy mailing lists use the unified search at:
> http://galaxyproject.org/search/mailinglists/
>
___________________________________________________________
Please keep all replies on the list by using "reply all"
in your mail client. To manage your subscriptions to this
and other Galaxy lists, please use the interface at:
http://lists.bx.psu.edu/
To search Galaxy mailing lists use the unified search at:
http://galaxyproject.org/search/mailinglists/
