[Date Prev][Date Next] [Chronological] [Thread] [Top]

Re: syncRepl fails to replicate new trees (ITS#2948)

--On Friday, March 05, 2004 9:35 AM -0800 Quanah Gibson-Mount 
<quanah@stanford.edu> wrote:

> Jong,
> I've now tested using a very tiny DB (just Stanford's basic setup, plus
> my entry so I can write to the master).  Everything works perfectly.
> There is one other difference -- I used a disk based BDB cache instead of
> a memory based BDB cache.  I'm now going to retry using the memory cache.
> If that works, I can only conclude that it is the large size of our DB
> that is causing syncRepl to fail.


It works with memory cached BDB as well (which I expected).

So, it looks like the shear magnitude in size of Stanford's DB is what is 
causing the syncRepl problems.


Quanah Gibson-Mount
Principal Software Developer
ITSS/TSS/Computing Systems
ITSS/TSS/Infrastructure Operations
Stanford University
GnuPG Public Key: http://www.stanford.edu/~quanah/pgp.html