Received: (qmail 9233 invoked by uid 2012); 29 Jun 1998 04:04:34 -0000 Message-Id: <19980629040434.9232.qmail@hyperreal.org> Date: 29 Jun 1998 04:04:34 -0000 From: Justin Hurley Reply-To: jjhurley@epo.net.au To: apbugs@hyperreal.org Subject: logs over filled with the same msg X-Send-Pr-Version: 3.2 >Number: 2522 >Category: mod_log-any >Synopsis: logs over filled with the same msg >Confidential: no >Severity: non-critical >Priority: medium >Responsible: apache >State: closed >Class: change-request >Submitter-Id: apache >Arrival-Date: Sun Jun 28 21:10:00 PDT 1998 >Last-Modified: Mon Jul 6 20:50:45 PDT 1998 >Originator: jjhurley@epo.net.au >Organization: >Release: 1.2.5 >Environment: OS/2, v4, fp6, current SSL distribution ApacheSSL-OS2-intl/1.2.5 SSLeay/0.8.1 Apache-SSL-OS2 (intl) 1.2.5 port by Antony T Curtis Original Apache for OS/2 port by Garey Smiley Howerver, request applies to all versions. >Description: refer PR 2217 (and others) There are several problems where user's logs files have filled up for various reasons, it would be a good thing if this could be avoided. (I know, there are various workarounds that can be implemented when a problem, usually a config problem, is indentified). But it would avoid a number of servers needlessly failing. (FWIW I am regularly suffering such symptoms in my test environment, but I believe it is a problem with a my config, so I'll do more testing before posting my problem about the following message which is only stoppable through a reinit.cmd) [Sun Jun 28 00:03:59 1998] select: (listen): Socket operation on non-socket >How-To-Repeat: >Fix: When an item is written to an error log, the last and 2nd last written entry should be checked. If this error is the same, a 'special' entry should be written, saying last msg repeated n times. This line could then be updated everytime the message is immediately repeated. Such an entry would stop logs from over-flowing with the same message looped. >Audit-Trail: Release-Changed-From-To: 1.2.5 SSL0.8.1-1.2.5 Release-Changed-By: coar Release-Changed-When: Thu Jul 2 03:46:07 PDT 1998 State-Changed-From-To: open-closed State-Changed-By: dgaudet State-Changed-When: Mon Jul 6 20:50:45 PDT 1998 State-Changed-Why: Upgrade to 1.2.6, this is one of the things that was dealt with. From src/CHANGES: *) Increase the robustness of the child_main loop. When unexpected select() or accept() errors occur we exit() the child. This deals with many reported problems where apache would fill the error_log with messages. [Dean Gaudet] PR#1747, 1107, 588, 1787, 987, 588 But there's no comparisons to reduce log messages -- given that any number of children can be logging, there's not a heck of a lot we can do. If you want that, use piped logs and filter there. Dean >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. ]