Mailing List ArchiveSupport open source code!
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]Re: Apache rotate log config option is where?
- To: tlug@example.com
- Subject: Re: Apache rotate log config option is where?
- From: Jake Morrison <jake_morrison@example.com>
- Date: Mon, 28 May 2001 04:06:58 -0700 (PDT)
- Content-Type: text/plain; charset=us-ascii
- In-Reply-To: <F57h05YFRcmVFOh1vdk00014ff9@example.com>
- Reply-To: Jacob.Morrison@example.com
- Resent-From: tlug@example.com
- Resent-Message-ID: <gWJMND.A.dX.WFjE7@example.com>
- Resent-Sender: tlug-request@example.com
Jc, Under Unix, a "zombie" has a specific meaning. It's a child process that has exited, and the system is holding its return status around, but there is no parent process left around that cares to read it. They still show up on the ps display, but there is nothing there. So they are the "living dead". In your case, it sounds like the processes are hung, waiting for something to happen, or they are just working slowly. Sometimes this happens when all the processes are waiting on a single resource, like an overloaded disk or network connection. You might check to see if there is any data actually moving, which will tell you which program is really causing the problem, client or httpd. The "strace" utility allows you to connect to a running process and trace its use of system calls. If you see reads and/or writes, data is moving. I think you would be better off with piping the logs through rotatelogs, as Tom does. Jake --- Jean-Christian Imbeault <jean_christian@example.com> wrote: > >I would hesitate to hack logrotate, as it may be used for other > stuff (e.g. rolling system logs). > > I would be hacking logrotate per se. Just making a seperate conf file > for apache where it would be told to use SIGUSR1 instead of HUP. > > >Sounds like poor error handling in the client programs. > > Agreed. > > >Are the clients in Perl? The error handling in the Perl libnet > >library has traditionally been pretty weak. > > Don't know. The clients are Net Appliance NetCache cache servers. > > >The httpd children can also wait around quite a while before > >timing out on a bad connection (10 minutes, say). > > In this case they actually don't ever time out ... I just end up > with a zombie. (I use the term zombie but they do respond properly > to a kill command). I end up with four things actually: > > - the perl script I use to tell the client to start requestuing > files using HTTP GETs one by one ... it's still in the process > list but obviously somethings wrong. > - a hung client ... > - one or more httpd children I assume belong to the connection > request by the now hung client > - a cron job waiting to send me the result of the output of my perl > script which hasn't finished. When I kill the perl script I get > my output file emailed, but of course it is incomplete. > > >Having lots of slow clients can really consume resources on a web > >server, as each one requires a httpd for the duration. > > So true ... > > Jc > _________________________________________________________________________ > Get Your Private, Free E-mail from MSN Hotmail at > http://www.hotmail.com. > > ----------------------------------------------------------------------- > Next Technical Meeting: Sat, May 12 13:30- > Next Nomikai Meeting: Fri, June (TBA) 19:30- Tengu Tokyo Eki Mae > ----------------------------------------------------------------------- > more info: http://www.tlug.gr.jp Sponsor: Global Online > Japan > __________________________________________________ Do You Yahoo!? Yahoo! Auctions - buy the things you want at great prices http://auctions.yahoo.com/
- References:
- Re: Apache rotate log config option is where?
- From: "Jean-Christian Imbeault" <jean_christian@example.com>
Home | Main Index | Thread Index
- Prev by Date: Re: Apache question
- Next by Date: Re: Apache question
- Prev by thread: Re: Apache rotate log config option is where?
- Next by thread: Apache question
- Index(es):
Home Page Mailing List Linux and Japan TLUG Members Links