httpd-bugs mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From bugzi...@apache.org
Subject DO NOT REPLY [Bug 26799] New: - httpd process grows rapidly to > 1GB until server runs out of swap
Date Mon, 09 Feb 2004 18:45:35 GMT
DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG 
RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT
<http://nagoya.apache.org/bugzilla/show_bug.cgi?id=26799>.
ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND 
INSERTED IN THE BUG DATABASE.

http://nagoya.apache.org/bugzilla/show_bug.cgi?id=26799

httpd process grows rapidly to > 1GB until server runs out of swap

           Summary: httpd process grows rapidly to > 1GB until server runs
                    out of swap
           Product: Apache httpd-2.0
           Version: 2.0.48
          Platform: Sun
        OS/Version: Solaris
            Status: NEW
          Severity: Major
          Priority: Other
         Component: Core
        AssignedTo: bugs@httpd.apache.org
        ReportedBy: glenn@apache.org


This bug has been seen on two different servers, with two different
versions of apache 2 using the worker MPM.

Sun Netra (1 CPU), Solaris 8, apache 2.0.47, occurred 5 times over last 3 months.
Sun E450 (4 CPU), Solaris 7, apache 2.0.48, occurred 3 times over last week.

On both of the above systems we have seen an apache httpd process rapidly
(in a few minutes) grow to > 1 GB in size until the system swap is exhausted
and processes start failing due to low memory.

This is a fairly standard install of Apache 2.0 using the worker MPM.
SSI gets used a great deal on these systems.  There are two non standard
apache modules installed, mod_jk 1.2.5 and a thread safe version of
mod_auth_mysql for apache 2.  The problem has been seen in instances of
apache which is not using mod_jk.

When this has occurred I have truss'd the process.  Trussing the process
results in no output so when the httpd process runs away like this it is
not using sytem io calls.

We have generated core files for the runaway processes using gcore but
Sun's dbx could not analyze the core file, it says they are corrupt.

We are trying to catch this problem when it occurs early enough so we
can get a good core file to analyze but have not been able to yet.

Any tips on how to debug this problem would be appreciated.

Thanks,

Glenn

---------------------------------------------------------------------
To unsubscribe, e-mail: bugs-unsubscribe@httpd.apache.org
For additional commands, e-mail: bugs-help@httpd.apache.org


Mime
View raw message