Matus UHLAR - fantomas wrote:
Hmm, is squid still unable to work if one of cache dirs has problems?
sounds like calling for bug report ;)
On 26.09.08 17:52, Amos Jeffries wrote:
It's already reported long ago. Made it onto the worklist for Squid-3
recently. Should be done Someday Soon Now (tm) :-).
If you mean bug 410, I wasn't sure if it's the same... quite possibly it is
but for full cache_dir (which the bug is about) there may be other ways to
handle it...
I did. A great solution is still being looked for.
Single object read failures are already recoverable (object is erased
and replaced).
The write failures are currently fatal as it happens in parallel to
network on large objects. And particularly bad if its a disk-full fatal
like the bug was about.
Amos
--
Please use Squid 2.7.STABLE4 or 3.0.STABLE9
