Received: (qmail 15715 invoked by uid 2012); 21 May 1999 18:28:32 -0000 Message-Id: <19990521182832.15714.qmail@hyperreal.org> Date: 21 May 1999 18:28:32 -0000 From: Konstantin Morshnev Reply-To: moko@design.ru To: apbugs@hyperreal.org Subject: F_SETLKW: Error, then server brings Linux down X-Send-Pr-Version: 3.2 >Number: 4453 >Category: os-linux >Synopsis: F_SETLKW: Error, then server brings Linux down >Confidential: no >Severity: serious >Priority: medium >Responsible: apache >State: open >Class: sw-bug >Submitter-Id: apache >Arrival-Date: Fri May 21 11:30:01 PDT 1999 >Last-Modified: Wed May 26 18:20:01 PDT 1999 >Originator: moko@design.ru >Organization: >Release: 1.3.6 >Environment: RedHat 5.2 Linux 2.0.36 #1 Tue Dec 29 13:11:13 EST 1998 i686 unknown default kernel, default kernel config no NFS, no custom modules in apache >Description: Server has relatively big load (evarage rate is about ~20 hits/sec). Today I've got following in error_log: ----------- [Fri May 21 15:46:25 1999] [emerg] (14)Bad address: fcntl: F_SETLKW: Error getting accept lock, exiting! Perhaps you need to use the LockFile directive to place your lock file on a local disk! [Fri May 21 15:46:26 1999] [alert] Child 21893 returned a Fatal error... Apache is exiting! [Fri May 21 15:50:30 1999] [error] (32)Broken pipe: accept: (client socket) [Fri May 21 15:50:32 1999] [error] (32)Broken pipe: accept: (client socket) ... ----------- Since that moment server began to slowdown of request processing, but load increased. This server has MaxClients 200, and I often see that all child are busy, but even in that situation there is no swaping and no big load. At 16:11 Linux gone down (or may be in swap?). Uptime was about a month, and I don't see any other reasons for server to go down. >How-To-Repeat: I don't know. >Fix: In fact, I think that error message is Linux bug, but all that came after it is sertanly apache bug. I do think it is nessesary to test and correct Apache behavour in case of such error. >Audit-Trail: Severity-Changed-From-To: critical-serious Severity-Changed-By: coar Severity-Changed-When: Wed May 26 08:09:00 PDT 1999 Category-Changed-From-To: general-os-linux Category-Changed-By: coar Category-Changed-When: Wed May 26 08:09:00 PDT 1999 From: Konstantin Morshnev To: coar@apache.org, apbugs@Apache.Org Cc: Subject: os-linux/4453: F_SETLKW: Error, then server brings Linux down Date: Thu, 27 May 1999 05:11:57 +0400 coar@apache.org wrote: > Synopsis: F_SETLKW: Error, then server brings Linux down > > Severity-Changed-From-To: critical-serious > Severity-Changed-By: coar > Severity-Changed-When: Wed May 26 08:09:00 PDT 1999 > Category-Changed-From-To: general-os-linux > Category-Changed-By: coar > Category-Changed-When: Wed May 26 08:09:00 PDT 1999 I'm not sure that your classification is correct. I think error is in Apache behavour in case of this error (that in brings OS down). This behavour identical for all OS, not only Linux. If you what, you can simulate this error in any OS and trace how Apache works. Sorry, I don't well understand this part of Apache to test it myself. WBR, MoKo >Unformatted: [In order for any reply to be added to the PR database, ] [you need to include in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ] [If you do not include this Cc, your reply may be ig- ] [nored unless you are responding to an explicit request ] [from a developer. ] [Reply only with text; DO NOT SEND ATTACHMENTS! ]