[olug] Argh!

Sam Tetherow tetherow at nol.org
Tue Dec 17 05:20:19 UTC 2002


And that would achieve?

Brian Roberson wrote:

>one word...
>
>ulimit
>
>
>
>
>----- Original Message -----
>From: "Sam Tetherow" <tetherow at nol.org>
>To: <olug at olug.org>
>Sent: Sunday, December 15, 2002 11:24 PM
>Subject: Re: [olug] Argh!
>
>
>  
>
>>You would have to check the source code for mozilla or gtfp but I would
>>bet what the author does is open the file for writing when they go to
>>save which will cause the file to be wiped.
>>
>>As for full memory and swap, many things write to log files or syslog
>>and most programmers do not take the time to handle this gracefully.
>> When say your /var partition fills up and syslog can no longer write to
>>file the os will buffer the writes in memory in the vain hope that some
>>disk space free up, when this doesn't occur, memory fills up, then swap
>>space, once swap is full the writes start failing, as well as any new
>>memory allocation.  Most programmers do not take the time to gracefully
>>handle a failed alloc beyond printf(stderr, "cannot allocation
>>memory\n"); exit;  But even those that do what is the proper behavior?
>> How will apache serve a file if it cannot allocate the memory to read
>>the file?    How is mysql suppose to handle a request if it cannot
>>malloc memory for the temporary buffers it needs to build the results set?
>>
>>    
>>






More information about the OLUG mailing list