Hi Ron, I posted this to the mailing list so that all can see the thread. I agree with your comments below, leave it to you and Don Dossa to work out the appropriate configuration. Thanks!
Jeff > From: Ronald Lambert <[EMAIL PROTECTED]> > Date: Tue, 1 Aug 2006 12:52:15 -0400 > To: Jeffrey P Kantor <[EMAIL PROTECTED]> > Cc: Don Dossa <[EMAIL PROTECTED]> > Subject: Re: data realibility > > Jeff, > To continue the discussion. Presumably the second disk set is > connected to the same CPUs or is there a separate complete set of cpu > +disk for the second copy? That would obviate faults from a cpu/ > motherboard failure writing garbage to both image disk sets. In turn > each set of cpu+disk could be supplied from different electrical > phases to ensure power integrity. > Would you agree that this is almost our most important point of > failure; the summit. If we lose a data image here it can never be > recovered for that point in time. So it almost seems like we should > be aiming for 10yr*365dy*24hr uptime at this section of the > operation. Alerts and transfers can always take place a day late or > whatever. > > ron > On Aug 1, 2006, at 11:22 AM, Jeffrey P Kantor wrote: > >> Hi Ron, >> >> Yes it is now two sets, but only temporarily on the mountain. Per our >> TechAssessWG, we can keep two backup sets on the mountain while we are >> sending one to the base. We keep the first mountain copy until we >> confirm >> that the data has reached the Base. Then we can delete one of the >> Mountain >> copies. When the data has reached the Archive Center, we have 2 >> copies at >> base and archive, so we can delete the second mountain backup. >> This should >> be within 24 hours, but could be as long as 48 hours, per our MTTRs. >> Similarly, once we have confirmed that the data has reached the >> Data Access >> Center, we then have 2 copies, at the Archive Center and DAC, so we >> can >> delete the Base copy. >> >> So, it sounds like we need 4 nights worth of data storage for >> backup on the >> mountain (2 copies times 2 nights buffer). >> >> Jeff >> >> >>> From: Ronald Lambert <[EMAIL PROTECTED]> >>> Date: Mon, 31 Jul 2006 14:53:48 -0400 >>> To: Don Dossa <[EMAIL PROTECTED]>, jeff kantor <[EMAIL PROTECTED]> >>> Subject: data realibility >>> >>> Don/Jeff, >>> Given the recent discussions I think we need a minimum of two sets of >>> the data on the mountain at any one time. This protects against all >>> the possible failures downstream. >>> ron >>> >> > > _______________________________________________ LSST-data mailing list [email protected] http://www.lsstmail.org/mailman/listinfo/lsst-data
