On 20-Apr-13 17:00, Tanstaafl wrote:
Another question - are there any caveats as to which filesystem to use for a mail server, for virtualized systems? Ir do the same issues/questions apply (ie, does the fact that it is virtualized not change anything)?
Problem of virtualized filesystem is not that it is virtualized, but that it is located on datastore with more virtual systems, all of them competing for the same i/o. *That* is the bottleneck. If you switch reiser for xfs or btrfs, you might win (or loose) a few %. If you optimize your esxi-datastore design, you might win much more than what you have ever dreamed of. I have 8 VMs (out of them 6 are Gentoo) hosted on ESXi, intended for various tasks (mail, dns, mysql, web, etc), moderately loaded. I used hw-raid controller with 2x sata-hdd in raid1 but performance was quite dissapointing and I experienced all sorts of i/o jams. Then I switched hdd for ssd (yes I use 2 of them in raid1, even if this is not generally recommended) and performance rocks now! I can start now kernel compilation on all 6 VMs at the same time, with near-zero performance penalty (depending on cpu/vcpu ratio and number of threads used). Unthinkable with hdd-based datastore. I would definitely recommend using SSD. Either directly as datastore for VMs, or at least as EXSi host-cache. There is also possibility of "hybrid-raid" (1xSSD and 1xHDD in raid1) on some raid-controllers. Or if your pocket is really deep, you could grab one of those FusionIO-cards to avoid being limited by rather slow sata-interface (SSD for PCIe)... Jarry -- _______________________________________________________________ This mailbox accepts e-mails only from selected mailing-lists! Everything else is considered to be spam and therefore deleted.
