The values are partially arbitrarily and a bit for reason chosen. The reason is (or at least was at the time of choosing them, that the active path group is selected as the one that has the highest priority count (all priority values of all paths in the group added together). So if the values 3 to 0 would have been used, 4 suboptimal paths would be used if only two optimal ones were available (4x2=8 > 2x3=6). At the time of coding the values I assumed 2 optimal paths should still be better than the sub-optimal paths. What would be the expectation here? I guess if the need is there it shouldn't be that hard to add configurable values as an argument to the command. Regards, Stefan Bader SW Linux on zSeries Development & Services Stefan.Bader@xxxxxxxxxx ---------------------------------------------------------------------------------- When all other means of communication fail, try words. dm-devel-bounces@xxxxxxxxxx wrote on 04.10.2006 20:24:07: > Are the relationships between various dm-multipath ALUA assigned path > priorities (active/optimal=50;active/sub-optimal=10;standby=1;other=0) > based on the performance characteristics of a particular array or were > they arbitrarily chosen? Also, is it reasonable to make these ALUA > assigned path priorities configurable (particularly the ratio of > active/optimal to active/sub-optimal priorities) on a per storage system > basis? > > -- > dm-devel mailing list > dm-devel@xxxxxxxxxx > https://www.redhat.com/mailman/listinfo/dm-devel -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
