ID: 15198 Updated by: [EMAIL PROTECTED] Reported By: [EMAIL PROTECTED] Status: Analyzed Bug Type: OCI8 related Operating System: Redhat Linux 6.1 PHP Version: 4.1.1 Assigned To: thies New Comment:
I've also got a PHP application that suffered from the OCI8 Recursive call! error after an upgrade to PHP 4.1.1. It worked fine with PHP 4.0.x, but with 4.1.x it randomly chokes with that error. There is nothing special going on. I'm not using bindings, I've just got scripts that logon, parse, execute, insert, delete, logoff, etc. Because of this, I'm forced to stick with the 4.0.x tree. I'd give you more debugging information, but it is a deployed system and I don't want to introduce the bugs into it. Previous Comments: ------------------------------------------------------------------------ [2002-02-05 05:07:25] [EMAIL PROTECTED] are you sure that you are not hitting the time_limit? reverting to 4.0.6 is not a smartthimg (tm) to do as a recursive oci call might kill your oracle MTS (if you use it). that's the only reason this catch got implemented. maybe useing ociinternaldebug(1) and sending me the output (and just the output of the oci module) would help. ------------------------------------------------------------------------ [2002-02-05 03:30:24] [EMAIL PROTECTED] Well, I have investigated the situation a bit more: I have a script which takes a long time (a maintenance script that runs at night, using php cgi binary), so I have set set_time_limit(7200) (which is two hours!) In the script, there's a complex query that takes 3 minutes to run. During this query, I get this OCI8 Recursive Call error and the script fails, even though the time limit has not been reached yet. Is there some internal timeout in the OCI calls? I now have to downgrade to 4.0.6 to prevent the problem. ------------------------------------------------------------------------ [2002-01-25 10:40:07] [EMAIL PROTECTED] this happens when an oci-call gets interrupted by a signal. this should only happen in 2 stituations: - you script hits max_execution_time (bad) - you did an apachectrl restart instead of apachectrl graceful the message "recursive call" is consitered a fatal-error and the apache child that generated that message will be terminated by the php oci-driveri. this is cause i've seen oracle MTS servers crashing when a client tried to call the oci* stuff in a reentrant way. i currently know no better way to save my "unbreakable" oracle from crashing;-) ------------------------------------------------------------------------ [2002-01-25 10:39:26] [EMAIL PROTECTED] this happens when an oci-call gets interrupted by a signal. this should only happen in 2 stituations: - you script hits max_execution_time (bad) - you did an apachectrl restart instead of apachectrl graceful the message "recursive call" is consitered a fatal-error and the apache child that generated that message will be terminated by the php oci-driveri. this is cause i've seen oracle MTS servers crashing when a client tried to call the oci* stuff in a reentrant way. i currently know no better way to save my "unbreakable" oracle from crashing;-) ------------------------------------------------------------------------ [2002-01-24 04:44:03] [EMAIL PROTECTED] I have a script that does an OCILogin, then OCIParses a query, then does a few OCIBindByName calls, an OCIExecute, an OCIFreeStatement and finally OCILogoff. The query is a PL/SQL block with a BEGIN, a few update and insert queries and an END. Most of the time the script runs fine. But every now and then (haven't discovered any regularity yet) I'm getting 'OCI8 Recursive call!' in my error logs. I've taken a look at ext/oci8/oci8.c to see what would trigger this error, and from what I see there, I can think of no errors in my script that could cause this. So my guess os that it is some internal error. ------------------------------------------------------------------------ -- Edit this bug report at http://bugs.php.net/?id=15198&edit=1
