Jeff Dike wrote: > [EMAIL PROTECTED] said: >> In one problem case, the dstat on the host shows a constant read >> and write load, with the read being perhaps a fourth of the amount >> written. The write load and the total throughput inside UML is >> stable somewhere between 300 to 600 kB/s, depending on the drive >> tested. This would look like it is doing partial writes all the >> time, requiring it to read as well as write. >> >> In the other problem case, the dstat on the host shows only write >> load of good magnitude, around 16 MB/s. But it is writing a whole >> lot more than what happens on the guest side. In fact, in one case, >> 15 times more (7.5 GB vs. 500 MB). The throughput on the guest side >> is a little better, 1.5 MB/s to 3.0 MB/s depending on the drive, >> but the drive is screaming like mad when this is happening. This >> would look like it is rewriting the same thing over and over again. > > Both of these are interesting. I don't know why UML would either be > doing lots of partial writes or multiple writes. > > I have been adding AIO support to the driver, and I discovered that > the driver is breaking up large I/Os on 2.6. The way cow support is > implemented, the driver will only send a page to the disk at a time. > However, the 2.6 block layer is sending up to 64K (that I've seen) > chunks, which are being broken up. > > This doesn't seem related to either of the problems you're > describing, though. > > The loopback device thing is interesting. I don't know what it > means. Theoretically, the ubd driver should be acting as a loopback > device itself.
We are lacking resources to investigate this further at the moment. What we have found out though is this: The above mentioned problems appear only when the actual disk image is sparse (eg. like uml_moo does them, or just created like that) or when using COW files (regardless if the backing file is sparse or not). When the disk image is flat (non-sparse), the performance is good and no oddities show in dstat. And the loopback device fixes the problem for sparse files (obviously testing loopback with cow files is more or less impossible). Also, using a 2.4 guest kernel seemed to not exhibit this particular problem at all - but we didn't have time to test it fully to give conclusive reports. -- Naked ------------------------------------------------------- SF email is sponsored by - The IT Product Guide Read honest & candid reviews on hundreds of IT Products from real users. Discover which products truly live up to the hype. Start reading now. http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click _______________________________________________ User-mode-linux-user mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/user-mode-linux-user
