FYI - You can run Veritas DMP on top of mpxio devices. DMP just sees a device of one path and goes about its business.
Jason King wrote: > Unofrtunately no, our standard is Veritas DMP. > > On 10/17/06, *yu larry liu* <[EMAIL PROTECTED] > <mailto:[EMAIL PROTECTED]>> wrote: > > Jason, > > One question for you. Is MPXIO enabled under your circumstance? > > If MPXIO enabled, each lun should only be shown once when running > format. > > Thanks, > Larry > > Roger Dong 已写入: > > > Thanks for the information. We will work on this issue and keep you > > updated. > > > > Best Regards, > > Roger > > > > Jason King 写道: > > > >> I don't know if I'll be able to redo the setup to get more details, > >> but basically it was most obvious with a large number of luns > >> (several hundred) + veritas DMP + multiple paths to a clariion > array. > >> Due to our setup, each lun on the Clariion showed up twice (one for > >> each controller) on every path to the array. Since DMP was being > >> used, this meant it looked like several thousand luns (a few > hundred > >> luns x 2 controllers x 4-5 paths to the luns) was visible to the > >> system. Also, all the paths that went through the standby > controller > >> showed up as offline, which probably contributed to the effect. > >> > >> On 10/15/06, *Roger Dong* <[EMAIL PROTECTED] > <mailto:[EMAIL PROTECTED]> > >> <mailto: [EMAIL PROTECTED] <mailto:[EMAIL PROTECTED]>>> wrote: > >> > >> Hi Jason, > >> > >> How many disk/lun in your system which take two hours to lunch > >> 'format'? > >> > >> We are aware of one performance issue in format. 'format' check > >> disk on-line status each time after S10U1. It will cost 2 minutes > >> for each disk/lun. We are investigate to improve this now. > >> > >> We are not aware of cfgadm issue. We will investigate it and > >> update you later. > >> > >> Thanks, > >> Roger > >> > >>> > >>>> *From: * Jason King <[EMAIL PROTECTED] > <mailto:[EMAIL PROTECTED]> > >>>> <mailto:[EMAIL PROTECTED] > <mailto:[EMAIL PROTECTED]>>> > >>>> *Date: * October 11, 2006 9:01:24 PM MDT > >>>> *To: * [email protected] > <mailto:[email protected]> > >>>> <mailto:[email protected] > <mailto:[email protected]>> > >>>> *Subject: * *[storage-discuss] cfgadm & format performance* > >>>> > >>>> Unfortunately, I have no means to see if this is present in > >>>> Nevada, but on Solaris 10U1, we've encountered some issues where > >>>> cfgadm and/or format seems to suffer poor performance when the > >>>> number of luns and/or controllers increase. > >>>> > >>>> In one extreme case, it takes over two hours from the time > >>>> 'format' is launched to when it actually displays the list of > >>>> devices. cfgadm -al also takes a rather absurd amount of time. > >>>> > >>>> Are there any plans to possibly improve the responsiveness of > >>>> these commands? > >>>> > >>>> > >>>> This message posted from opensolaris.org > <http://opensolaris.org> <http://opensolaris.org> > >>>> _______________________________________________ > >>>> storage-discuss mailing list > >>>> [email protected] > <mailto:[email protected]> > >>>> <mailto: [email protected] > <mailto:[email protected]>> > >>>> http://mail.opensolaris.org/mailman/listinfo/storage-discuss > >>>> < http://mail.opensolaris.org/mailman/listinfo/storage-discuss> > >>> > >> > >> > > > > > ------------------------------------------------------------------------ > > _______________________________________________ > storage-discuss mailing list > [email protected] > http://mail.opensolaris.org/mailman/listinfo/storage-discuss > _______________________________________________ storage-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/storage-discuss
