>Number: 2016 >Category: os-solaris >Synopsis: Under some circumstances files 'break' - and remain so. >Confidential: no >Severity: serious >Priority: medium >Responsible: apache >State: open >Class: sw-bug >Submitter-Id: apache >Arrival-Date: Mon Mar 30 09:10:01 PST 1998 >Last-Modified: >Originator: [EMAIL PROTECTED] >Organization: apache >Release: 1.3b5 >Environment: Solaris 2.4, unpatched. gcc 2.7.2.x >Description: In this case, a GIF file, which is certainly 'OK' in general fails to download from a virtual server. It is not a client issue - direct telnet confirms the problem.
The identical file will download fine from a different virtual host on the same server. Although the configs are not identical (of the two servers) making them so does not change the problem. Restarting the web server process does not fix the problem. Sometimes, but not always, copying the file will fix the problem. It is as if the file is 'locked' open (mmap()ed?). The server will always report a succesful (status 200) access. The client will hang on 'host contacted, waiting for reply' I have extensively investigated this problem, and it's either very stupid, or very bizarre. Although NFS is involved in our setup, I am confident that is not relevant (I can replicate the problem on a non-NFS set up). Although I have not tried to test it on another OS, I'm guessing that this is a solaris problem. >How-To-Repeat: The problem *is* rather erratic. It can occasionally be made to go away by copying files. Currently, I have an example running as: http://www.erik.co.uk/TNGirl.gif http://www.eurotalk.co.uk/TNGirl.gif You will, I hope, find that the one under 'erik' works fine, the one under 'eurotalk' doesn't. >Fix: Looks to me like something freaky to do with overrunning buffers, or similar >Audit-Trail: >Unformatted: [In order for any reply to be added to the PR database, ] [you need to include <[EMAIL PROTECTED]> in the Cc line ] [and leave the subject line UNCHANGED. This is not done] [automatically because of the potential for mail loops. ]
