> Just for kicks, I wrote a little shell script to do 3 svnadmin load's (very
> write-centric) and 3 svnadmin verify's (very read-centric) on the various
> combinations of BDB vs. FSFS and local-disk vs. NFS-hosted. Here are the
> results (as averages ... the variations were small enough to be insignificant):
>
> svnadmin load of the first 500 revisions of The Subversion Source
> Code Repository (from a local-disk-stored dumpfile):
>
> BDB FSFS
> +---------+---------+
> NFS | 48.101s | 99.957s |
> +---------+---------+
> LOCAL | 71.676s | 65.633s |
> +---------+---------+
Michael and I were chatting on IRC about the above results. svnadmin
has an option to eliminate syncing data to disk after each commit
(syncing after the accessor is being closed). Additional tests on (on
his machine) with this option:
svnadmin load of the first 500 revisions of The Subversion Source
Code Repository (from a local-disk-stored dumpfile, with
the --bdb-txn-nosync flag for BDB). Old FSFS number provided
for comparison.
BDB FSFS
+---------+---------+
NFS | 39.226s | 99.957s |
+---------+---------+
LOCAL | 47.263s | 65.633s |
+---------+---------+
> svnadmin verify of the repository freshly-loaded with those first
> 500 revisions:
>
> BDB FSFS
> +---------+---------+
> NFS | 8.369s | 24.183s |
> +---------+---------+
> LOCAL | 8.045s | 6.659s |
> +---------+---------+
>
> Do what you will with these numbers. (I admit being shocked by the BDB-NFS
> vs. BDB-LOCAL "load" numbers ... maybe reading the dumpfile and loading into
> a repos on the same local hard-drive causes too much disk churn?)
HTH,
Erik.
---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@subversion.tigris.org
For additional commands, e-mail: users-help@subversion.tigris.org
Received on Thu Mar 8 22:26:40 2007