[OpenAFS-devel] enhance audit logs to support Sys V message queues

[Jason Edgecombe]

Steven Jenkins wrote:
> On Wed, Apr 22, 2009 at 9:20 PM, Russ Allbery <rra@stanford.edu> wrote:
>   
>> Steven Jenkins <steven.jenkins@gmail.com> writes:
>>
>>     
>>> I just uploaded a preliminary patch to openafs-bugs (RT #124674) that
>>> adds support to audit logs for Sys V message queues.  In other words,
>>> if you run this patch, instead of audit logs going to a file, they can
>>> go to Sys V message queues.  There is an additional option to
>>> configure if you want to experiment with this: --with-audit-interface,
>>> with two options: file and sysvmq, defaulting to file.
>>>       
>> Out of curiosity, what do you get from message queues that you don't get
>> from a named pipe?  No blocking if you have no reader?
>>
>>     
>
> You can choose to do non-blocking for named pipes.   However, named
> pipes (aka 'FIFO's to many in the Linux world) have a much tighter
> coupling between producers and consumers than message queues do.  For
> example, if a writer tries to write to a named pipe that  has no
> readers, a SIGPIPE (or EPIPE if SIGPIPE is being ignored) is
> generated
Sometimes, I need to get more debugging info after the fact. I'm just 
wondering if queues function as a circular buffer when no readers are 
present.  I'm wondering how useful it would be to be able to get the 
last X messages of maximum verbosity from a circular buffer. If 
something of interest happens, then ask the service to dump the circular 
buffer somewhere.

Granted, this doesn't come up often, and I'm not sure how this would be 
better or worse than a memdump or something similar.

Just thinking out loud,
Jason