On Tue, 2 May 2006, Graham Leggett wrote:
The need-size-issue goes for retrievals as well.
If you are going to read from partially cached files, you need a "total
size" field as well as a flag to say "give up, this attempt at caching
failed"
Are there partially cached files? If I request the last 200 bytes of a
4.3GB DVD image, the bucket brigade contains the complete file... The
headers says ranges and all sorts of things but they don't match
what's cached.
What may be useful is a cache header with some metadata in it giving the
total size and a "download failed" flag, which goes in front of the
headers. The metadata can also contain the offset of the body.
I solved it with size in the body and a timeout mechanism, a "download
failed" flag doesn't cope with segfaults.
OK. It's attached. It has only had mild testing using the worker mpm
with mmap enabled, it needs a bit more testing and auditing before
trusting it too hard.
Note that this patch fixes a whole slew of other issues along the way,
the most notable ones being LFS on 32bit arch, don't eat all your
32bit memory/address space when caching a huge files, provide
r->filename so %f in LogFormat works, and other smaller issues.
Is it possibly to split the patch into separate fixes for each issue
(where practical)? It makes it easier to digest.
It's possible, but since I needed to hammer so hard at mod_disk_cache
to get it in the shape I wanted it I set out to first get the whole
thing working and then worry about breaking the patch into manageable
pieces. For example, by doing it all-incremental there would have been
a dozen or so disk format change-patches, and I really don't think you
would have wanted that :)
As said, this is a preliminary jumbo patch for those interested in how
we tackled the various problems involved (or those who love to take
bleeding edge code for a spin and watch it falling into pieces when
hitting a weird corner case ;).
Also the other fixes can be committed immediately/soon, depending on how
simple they are, which will simplify the final patch.
Yup. I'll update bug#39380 when we feel that we have a good solution.
/Nikke
--
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Niklas Edmundsson, Admin @ {acc,hpc2n}.umu.se | [EMAIL PROTECTED]
---------------------------------------------------------------------------
To err is Human. To blame someone else is politics.
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
