We'll check this.

BTW, generically, what's your experience data for the issue if using this mechanism. This might be very useful for reference.

Simon.


On Sun, Aug 23, 2009 at 12:18 AM, Sriram Rao <sriramsrao@gmail.com> wrote:
Simon,

The logcompactor is designed for this purpose: it takes a bunch of log
files, and compacts them to produce a new checkpoint. It is setup to
run once an hour.  Did you disable that?

Sriram

On Sat, Aug 22, 2009 at 7:27 AM, Moreframe<moreframe@gmail.com> wrote:
> Sriram, All,
> We did a test in order to know how long the Metaserver can get recovery
> under 1 million files. The result is quite bizarre, logs captured as below
> -
>
> 08-12-2009 14:53:34.807 INFO - (metaserver_main.cc:85) Starting
> metaserver...
>
> 08-13-2009 07:19:13.768 INFO - (metaserver_main.cc:99) Starting metaserver
> finish
>
> Totally replayed 117 log files, 1 million data files, spent 16:25:38.961
>
> Pls support the issue, we think this duration of namespace recovery is too
> long to afford.
>
> Simon.
>
> ------------------------------------------------------------------------------
> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day
> trial. Simplify your report design, integration and deployment - and focus
> on
> what you do best, core application coding. Discover what's new with
> Crystal Reports now.  http://p.sf.net/sfu/bobj-july
> _______________________________________________
> Kosmosfs-devel mailing list
> Kosmosfs-devel@lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/kosmosfs-devel
>
>