|  support |  documentation |  report a bug |  advanced search |  search howto |  statistics |  random bug |  login
Bug #19796 Oracle query in array, returned by function: memory leak (no more high CPU load)
Submitted: 2002-10-07 06:33 UTC Modified: 2002-10-27 19:13 UTC
From: gko at gko dot net Assigned:
Status: No Feedback Package: Performance problem
PHP Version: 4.2.3 OS: Digital Unix V4.0F
Private report: No CVE-ID: None
View Add Comment Developer Edit
Welcome! If you don't have a Git account, you can't do anything here.
You can add a comment by following this link or if you reported this bug, you can edit this bug over here.
Block user comment
Status: Assign to:
Bug Type:
From: gko at gko dot net
New email:
PHP Version: OS:


 [2002-10-07 06:33 UTC] gko at gko dot net
I need to generate some reports and I need to have some table's data as a whole as some columns of a row in this table and other tables reference other rows of this table. The PHP version is 4.0.4pl1, Apache is 1.3.14, and Oracle 8.1.7.

So basically, I do this inside a function:
while (ora_fetch($cursor)) {
  $key = ora_getcolumn($cursor, 0); 
  for ($j=0; $j<ora_numcols($cursor); $j++) {
    $col_name=ora_columnname($cursor, $j);
    $col_value = ora_getcolumn($cursor,$j);
    $ret_array[$key][$col_name] = $col_value;

Here are my problems:
  - after the end of the script execution, the CPU load (seen by top) goes up to 99.90% for several minutes (for 4000 entries of 20 columns, it lasts about 2 minutes); during the script execution and HTTP transfer, the CPU load is normal, once the file transfer is over, the load becomes 99.x %
  - if I do this on a just started Apache, a lot of memory is consumed (normal) but not released:

nobody    10159  0.0  0.1 14.8M 2.3M ??       S    14:42:45     0:00.25 /www/bin/httpd
nobody    10159 99.9  0.7 42.7M  29M ??       R    14:42:45     0:36.23 /www/bin/httpd
                ^^^^ CPU  ^^^^  ^^^^^RAM

  - I did some experiements and observed this:
     - if I do an exit in the returning function just before "return $ret_array", then the high CPU load does not occurs (I'm not sure about memory leak)
     - if now I do an exit just after calling the function, then the high CPU load occurs (I'm not sure about memory leak). It looks like if $ret_array "moves" out of the function, a high CPU load occurs... I have tried with references, unset, etc... before/after calling the function to help the garbage collection at the end of the script but it doesn't help, I always have the high CPU load.

    Do you have any idea of what's happening?


Add a Patch

Pull Requests

Add a Pull Request


AllCommentsChangesGit/SVN commitsRelated reports
 [2002-10-07 07:23 UTC]
Thank you for taking the time to report a problem with PHP.
Unfortunately you are not using a current version of PHP -- 
the problem might already be fixed. Please download a new
PHP version from

If you are able to reproduce the bug with one of the latest
versions of PHP, please change the PHP version on this bug report
to the version you tested and change the status back to "Open".
Again, thank you for your continued support of PHP.
 [2002-10-09 04:43 UTC] gko at gko dot net
I have compiled PHP 4.2.3 (was 4.0.4pl1) and tested my code: the high CPU load after code execution is no more there, but there is still memory leak:

nobody    19621  0.0  0.0 14.3M 560K ??       S    16:32:09     0:00.00 /www/bin/httpd

nobody    19621  0.0  0.9 50.3M  37M ??       S    16:32:09     0:33.91 /www/bin/httpd
                          ^^^^^  ^^^

If I comment out "$retval[$key][$col_name]= $value", then I have:
nobody    20871  0.0  0.0 14.8M 1.9M ??       S    17:24:49     0:00.03 /www/bin/httpd
 [2002-10-09 09:38 UTC]
Please try using this CVS snapshot:
For Windows:

 [2002-10-09 11:18 UTC] gko at gko dot net
Does this 'latest' version fix some items that could be related to my problems?
 [2002-10-09 13:11 UTC] gko at gko dot net
What I'm wondering is wheter the memory is going to be given back to the system if it needs it ? With 4.0.4pl1, the  VSZ/RSS numbers were above 150 M and never shrunk, which worried the system maintenance staff. I know such memory readings have to be carefully interpreted but it's still troubling...
 [2002-10-10 15:52 UTC] gko at gko dot net
With 4.2.3, the link with Oracle seems to have some problems as the server looks less responsive and with many queries, I have "arithmeric exceptions" in error_log of httpd so I'm going back to 4.0.4pl1 and its high CPU load...
 [2002-10-12 09:53 UTC]
Please try using this CVS snapshot:
For Windows:

 [2002-10-27 19:13 UTC]
No feedback was provided. The bug is being suspended because
we assume that you are no longer experiencing the problem.
If this is not the case and you are able to provide the
information that was requested earlier, please do so and
change the status of the bug back to "Open". Thank you.

PHP Copyright © 2001-2024 The PHP Group
All rights reserved.
Last updated: Wed May 22 19:01:31 2024 UTC