Thanks to Casper Dik and Bismark Espinoza for their answers. There were problems with the disk, the SCSI box and probably with the cable as well, so I simply went back to the old configuration. ----------------------------------------------------------------------- Your SCSI bus is likely too long; that you don't see speed degradation isn't that abnormal; one disk normally doesn't saturate the chain. It is possible that the bus was already running in degraded mode. You will certainly have performance problems when running multiple tests on multiple disks in parallel (a 20MB/s bus can sustain 5MB/s for each disk; now that your SCSI chain is too long/noisy; you can get only 1.25MB per second per disk. On Tue, 12 Mar 2002, Andras Micsik wrote: > Hi, > > I connected a Quantum Atlas 1. (XP32150) with a roline 50-68 cable to our > existing chain of SCSI disks on an Ultra 1 with Solaris 7. It uses fas > drive, and scsi_options was 1ff8 before adding new disk. Before this > scsiinfo said that disks were fast (bus speed cca. 10 MB). Now scsiinfo > says that disks are accessed at cca. 5 MB/s. > > Although the bus speed was cut down to 5MB, I can't measure any I/O speed > degradation. I made > > time dd if=<file-bigger-than-100MB> of=/dev/null > > before and after, and the sys times are nearly the same, real time is > somewhat bigger. > > My questions: > > Is this normal? (Because it seems illogical) > > Is there a possiblity that the cable (roline) causes the new > Quantum disk seen as not Fast SCSI disk? > > Thanks in advance, > > -- Andras Micsik micsik@sztaki.hu MTA SZTAKI Hungary http://www.sztaki.hu/~micsik _______________________________________________ sunmanagers mailing list sunmanagers@sunmanagers.org http://www.sunmanagers.org/mailman/listinfo/sunmanagersReceived on Fri Mar 22 10:06:33 2002
This archive was generated by hypermail 2.1.8 : Thu Mar 03 2016 - 06:42:37 EST