[svn.haxx.se] · SVN Dev · SVN Users · SVN Org · TSVN Dev · TSVN Users · Subclipse Dev · Subclipse Users · this month's index

Re: Regression in bindings? 1.7/1.8 vs 1.6

From: Philip Martin <philip.martin_at_wandisco.com>
Date: Tue, 26 Aug 2014 11:39:29 +0100

Ivan Zhakov <ivan_at_visualsvn.com> writes:

> On 12 August 2014 12:46, Philip Martin <philip.martin_at_wandisco.com> wrote:
>> Philip Martin <philip.martin_at_wandisco.com> writes:
>>
>>> The txn_dir_cache gets cleared when the pool associated with the
>>> txn_root is cleared
>>
>> The net result is that the txn_dir_cache doesn't really work with
>> apache/mod_dav_svn. Import a directory containing a large number of
>> small files and the rate at which files are added slows as the number of
>> file increases. For svnserve the rate remains more-or-less constant.
>> The final stage comverting the transaction into a revision is much
>> faster for svnserve than for apache/mod_dav_svn.
>>
> Hi Philp,
>
> Ouch, it looks like the serious problem that could lead invalid data
> committed to repository. Is it true? Did you already filed issue?

Cache invalidation has a couple of effects:

  - the cache doesn't work very well for the most common access pattern,
    i.e. mod_dav_svn doesn't scale as well as svnserve.

  - the cache causes out-of-date data to be written for some unusual
    access pattern

I wrote a regression test (r1617500) for the second part but I have not
raised an issue.

-- 
Philip Martin | Subversion Committer
WANdisco // *Non-Stop Data*
Received on 2014-08-26 12:40:00 CEST

This is an archived mail posted to the Subversion Dev mailing list.

This site is subject to the Apache Privacy Policy and the Apache Public Forum Archive Policy.