tg%gmplib.org@localhost (Torbjörn Granlund) writes: > Greg Troxel <gdt%ir.bbn.com@localhost> writes: > > I'm not sure why you want to experiment :-) Seriously, on a production > machine, I would recommend just making all files used for vnd > non-sparse. (If you want to grab another box to experiment on a home to > fix the bug, great, but I would decouple that from your production > setup.) > > Well, you misunderstand. :-) > > I am all for a robust production machine... Now it is not robust, and I > would like to understand why it is not robust. Determining the envelope > of the bug helps; I will then intelligently choose life outside of said > envelope. Sure, but the widely-reported experiences of pretty much everybody is that sparse vnd files are trouble, and that without them, netbsd/xen is very stable. > A nice side-effect is doing a contribution to making NetBSD better. True, but you can fix this bug without involving your colo hardware. > Another reason not to use sparse files is that creating files non-sparse > gets you pretty good locality of block allocation and I think this leads > to higher efficiency, but I don't have test data. > > I think this is true for rust disks, less so for SSDs (at least if you > remember to coax the NetBSD installer into not misaligning all > partitions...). That could be true. But there may be larger IO chunks anyway. This needs some data, for the glorious future when vnd works with sparse.
Attachment:
pgp7jn3PqNdGP.pgp
Description: PGP signature