Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.
Am Dienstag, den 03.04.2012, 10:07 -0700 schrieb Digimer: > On 04/03/2012 02:53 AM, Lukas Gradl wrote: > > As we've just space for one 3.5" HDD (the other bay is used by the > > Boot-SSD) I'm unable to install a raid5-setup. > > Three things; > > 1. RAID 5 will shorted the life of SSDs. I don't think so - but you've to be careful which SSD you choose (we use Intel 320 and have a successful setup with Crucial M4) and not to use them to their full capacity (we normally leave 15-20 % of the space unused) > 2. RAID 1+0 is faster than RAID 5, if you can afford the reduced capacity. As I wrote: We've just space for one 3.5" HDD - so we're able to install 2 2.5" Disks with a special Adapter - so neither RAID 5 (3 disks minimum) nor Raid 1+0 (4 discs minimum) is possible. > 2. cat /sys/block/sdb/queue/rotational > > If set to '1'; > > echo 0 > /sys/block/sdb/queue/rotational > > That made a significant improvement on my test bed (albeit with hardware > RAID 5 [LSI-9265 w/ SSDs and HP P410i w/ 10krpm SAS HDDs). For which setup? At the moment I've a SATA-Disk, so rotational=1 seems quite right. As I wrote before: At the moment there's no RAID-Controller. > > As for dual-primary/Protocol C; I've got nine clusters in production, > some under heavy load, hosting both Linux and Windows VMs. These operate > just fine, with proper config of the storage. For example; in some > cases, I use a couple separate RAID 1 arrays, in others I used RAID 5. > All use traditional platter drives (some SATA, most SAS). None of these > clusters are over the top, performance wise. -- -------------------------- software security networks Lukas Gradl <proxmox#ssn.at> Eduard-Bodem-Gasse 6 A - 6020 Innsbruck Tel: +43-512-214040-0 Fax: +43-512-214040-21 --------------------------