Reviewers assigned incorrectly in OJS 3 migration

Software Support
1

We are in the process of migrating our old OJS site to version 3.1.1-4 and found that a number of the active articles we manually inputted into the new dashboard (that didn’t migrate over automatically) have either 1) moved directly into a second round of review without any content being added to Round 1 once we send the article to review, and/or 2) had reviewers automatically/incorrectly assigned to them in Round 1 that had been previously assigned to old articles.

Will reviewers be automatically assigned to articles if we don’t assign them? Is there any way we can delete the first rounds in these manually uploaded articles since they are either blank or incorrectly connected to reviewers?

2

Hi @lbosc,

Reviewers should not be manually assigned to articles. That is not typical behaviour of OJS and might be a result of a problem with the migration.

It is not currently possible to delete the first round of review or revert to the first round of review, other than by removing the review round from the review_rounds table in the database. However, you should only do this for the review rounds that are blank, not the ones that have reviewers assigned to them.

Best,
Amanda Stevens
Public Knowledge Project Team

3

Hi @astevens,

Sorry for any confusion, but we didn’t manually assign reviewers. The issue is that when we click the “send to review tab” for our articles, it opens up to three rounds of review. The first one is usually blank. The second has reviewers listed as automatically assigned (without adding them in). These reviewers (and their reviews) are reviewers dating back to last year and have nothing to do with the article in question. The third round is the actual round with the file we intend to send to review and the reviewers area empty for us to choose our reviewers.

It’s fairly inconsistent and sometimes only two rounds open up (where the first is blank, or with reviewers) and other times the system works properly.

I just don’t want these past reviewers to be getting assignments that we have not requested.

Thanks,
Lauren

4

Hi @lbosc,

I’m sorry, I made a mistake in my last post - I meant to say “Reviewers should not be automatically assigned to articles.” They should, in fact, be manually assigned! If your site is automatically assigning reviewers to submissions or opening new rounds of review of its own accord, there is some kind of problem on your site. Do you see any errors in your error log?

Best,
Amanda Stevens
Public Knowledge Project Team

5

Hi Astevens,
I am the tech person doing the migration for Lauren Bosc’s site.
I’ve turned ON the “show_stacktrace” and this is the current logs that our site generates

Reviewer-errorlog

thanks for your helps
-PTran

6

Hi @Ptran,

These are all warnings, not errors, and unfortunately they don’t tell us anything useful.

Best,
Amanda Stevens
Public Knowledge Project Team

7

hi Astevens,
what type of errors you would want to see? Can you let me know what need to be set to see those errors?

many thanks
-P

8

hi Astevens,
Can you please let me know how to trouble shooting this issue? it is become critical now for us.

thanks very much.
-P

9

Hi all,

As @astevens wrote, OJS shouldn’t ever be assigning reviewers automatically, so I suspect there’s something unusual with your database, e.g. perhaps there is already data in the review rounds table that corresponds to the newly-minted submission IDs that incoming submissions are being assigned.

Regards,
Alec Smecher
Public Knowledge Project Team

10

hi Alec,
what do I look for in the review_rounds table? Can I show you show you my “review_rounds” table?

thanks
-P

11

Hi @Ptran,

I would suggest looking in your review_rounds and review_assignments tables for entries with submission_id values that don’t correspond to any value of submission_id in the submissions table.

Regards,
Alec Smecher
Public Knowledge Project Team

12

HI Alec,
Sorry for a long delay, It seems to me I can do one query that identifies rows matching .Before I run them, can you double check if this is what you are asking me to do?

SELECT review_rounds.* from review_rounds WHERE review_rounds.submission_id NOT IN (SELECT submission_id FROM submissions)

and another that does

SELECT review_assignments.* from review_assignments WHERE review_assignments.submission_id NOT IN (SELECT submission_id FROM

submissions)

Or do we need to find some kind of overlap between these relevant rows from review_rounds and review_assignments?

thanks
-PT

13

hi Alex,
I got some output from my database. Hope that you can give me some suggestions.

I analyzing overlap between review_assignments and review_rounds as per submission_id .

select count(*) from review_assignments;

340

select count(*) from review_assignments where review_assignments.submission_id NOT IN (select submission_id from review_rounds);

0

select count(*) from review_assignments where review_assignments.submission_id IN (select submission_id from review_rounds);

340

select count(*) from review_rounds;

434

select count(distinct(submission_id)) from review_rounds;

405

select count(*) from review_rounds where submission_id NOT IN (select submission_id from review_assignments);

252

select count(*) from review_rounds where submission_id IN (select submission_id from review_assignments);

182

select count(distinct(submission_id)) from review_rounds where submission_id NOT IN (select submission_id from review_assignments);

244

select count(distinct(submission_id)) from review_rounds where submission_id IN (select submission_id from review_assignments);

161

so every row in review_assignments has a submission_id matching a

submission_id found in review_rounds and submissions, but some

review_rounds have submission_ids that are not found in

review_assignments (but always found in submissions)

14

Anyone ? ?
Much appreciate for you input !

15

Hi @Ptran,

Sorry for the delay in responding. Is it possible that you’re running some custom code or modifications that were added to create automatic review assignments? This is very unusual behaviour.

Regards,
Alec Smecher
Public Knowledge Project Team

16

I’m sure that we don’t have any custom code or modifications that were added. Everything just straight from the original upgrade OJS source code. Can you perhaps connect to our test server to take a look? I tried to connect to OJS tech support but they are so busy until end of April. If you can please let me know your quote.
thanks