LOVD 2.0

Welcome to the LOVD 2.0 bug tracking system. Please note that this bug tracking system is not for LOVD 3.0!

If you have any issues, please read the documentation and Frequently Asked Questions.
Please note that you need to register before you can submit bugs or feature requests.

Information By default, all closed tasks are hidden from view.
Click here to see all tasks.

Tasklist

FS#17 - Impossible to import anything anymore, import script returns empty page

Attached to Project: LOVD 2.0
Opened by Ivo F.A.C. Fokkema (ifokkema) - Friday, 21 September 2007, 11:33 GMT+2
Last edited by Ivo F.A.C. Fokkema (ifokkema) - Thursday, 14 February 2008, 10:44 GMT+2
Task Type Bug Report
Category Backend / Core → Config area
Status Closed
Assigned To Ivo F.A.C. Fokkema (ifokkema)
Operating System All
Severity High
Priority Urgent
Reported Version 2.0-beta-08
Due in Version 2.0-04
Due Date 2008-02-29
Percent Complete 100%
Votes 0
Private No

Details

All of a sudden the import script fails to import anything anymore. No error is shown, just an empty page is sent and the browser asks whether to save or open the file, so the content-type header is not being sent either. It seems like the PHP thread is crashing. Files with no new entries are processed just fine, only when new entries are present, this problem appears.
This task depends upon

Closed by  Ivo F.A.C. Fokkema (ifokkema)
Monday, 03 March 2008, 09:27 GMT+2
Reason for closing:  Fixed
Comment by Ivo F.A.C. Fokkema (ifokkema) - Friday, 21 September 2007, 11:56 GMT+2
This is a memory problem. The import script currently loads all patient data into the memory, even if the patient data is not necessary. This is a CPU vs Memory consumption consideration; more CPU power will be used when patient data will be loaded in chunks. A temporary fix is to increase the maximum memory usage limit of PHP. A more constructive fix will be included in 2.0-beta-09, where the patient data will be loaded only just before it\'s needed, one by one. This will increase the time it takes to import a large file into a relatively small database, but will not consume so much memory especially with larger databases and small import files will be processed quickly.
Comment by Ivo F.A.C. Fokkema (ifokkema) - Thursday, 14 February 2008, 10:43 GMT+2
It seems this problem keeps occurring, but also because of variant and patient2variant (allele, pathogenicity, etc) data. LOVD 2.0-04 fixes this partially, but still runs into problems with big data files and low memory usage settings. Getting the memory usage down more would result in a (much) slower import script, which we do not consider a good choice.

Loading...