Re: fsfs revprop packing in f5 Re: Does fsfs revprop packing no longer allow usage of traditional backup software?

On Thu, Jul 7, 2011 at 8:39 AM, Daniel Shahaf <danielsh_at_elego.de> wrote:
> Philip Martin wrote on Thu, Jul 07, 2011 at 08:36:06 +0100:
>> Daniel Shahaf <danielsh_at_elego.de> writes:
>>
>> > The process to edit a revprop that had been packed would be:
>> >
>> > * grab write lock
>> > * prepare a new pack-file-with-inline-manifest
>> > * move-into-place, atomically replacing the previous pack file
>> > * ungrab write lock
>> >
>> > That is what guarantees cp(1) consistency that Hyrum mentions.
>>
>> Atomic replace is going to involve a retry loop on Windows, and so could
>> potentially take a long time.  Virus scanners on repository disks?
>>
>
> Yes, this is going to be a problem.

Virus scanners on repo disks on Windows are *always* going to be a
problem. There isn't a whole lot we can do about it (aside from the
retry loop).

I'll again stress that I expect prop editing of packed revprops to be
a rare operation (save for r0, which I've addresses elsewhere).
Please don't let theoretical worst-case behavior drive the
implementation.

-Hyrum
Received on 2011-07-07 15:45:38 CEST