|
![](/i/fill.gif) |
On Wed, 26 Oct 2011 13:43:47 -0700, Darren New wrote:
>> Once upon a time, I worked for a Fortune 50 company with several
>> thousand Windows servers. Informix was running on them, and there was
>> a memory leak. "Reboot the system" was the "fix", to the extent that
>> the reboot was scripted and scheduled to run nightly.
>
> Same with Linux set-top boxes I was working on. Because we didn't have
> the source to the code that was leaking the memory.
>
> And indeed, I was working on one Linux server system where they were
> catching sigsegv's and outputting "caught a signal!" to stdout, but then
> not actually recovering. So I had to pipe the output into a second
> process *I* wrote that would grep for that string and send a kill -9 and
> restart it when it got stuck.
>
> It's not Windows per se, but proprietary software that you can't fix.
Exactly. :)
>> Now, part of the reason for that was that Informix was taking their
>> time fixing the problem
>
> There ya go. :-)
Yep.
Jim
Post a reply to this message
|
![](/i/fill.gif) |