Re: How to handle "Internal Server Error"
Pavel Francirek <[email protected]>
| Newsgroups | gmane.comp.web.fastcgi.devel |
|---|---|
| Organization | Seznam.cz a.s. |
| Message-ID | <1122560494.30316.28.camel@cryptobook-acer2304> |
On Thu, 2005-07-28 at 13:05 +0000, jancha wrote: > Hi. > > On Thursday 28 July 2005 00:43, Jack wrote: > > Hello Jancha, > > > > Thanks a lot for the reply. It is very helpful. Two more questions: > > > > Re: 2) I found the 600 seconds definition in mod_fastcgi source files. > > I think I'll change that and rebuild mod_fastcgi. I plan to make it 0 or 1. > > Is there any reason I shouldn't make it so short? > > This can be done, however, might be a bad idea because: > if your program is faulty, fastcgi will keep re-launching it all the time - > and what if the reason for crashing is for instance some kind of memory leak? > it might be unsafe.. OK, I understand it. But why only 3 failures in 30 seconds? Sometime even 30 failures in 1 seconds is still better than complete stop for 10 minutes. Especially if there is webserver farm. Small failure may crash whole cluster for minutes. Pavel ___________________________________ fastcgi-developers mailing list http://fastcgi.com/fastcgi-developers/