Thunderbird crashing just as I hit send on my reply yesterday probably made it disappear.
Hopefully they didn't do something like use WD Purple drives. Had an array with a failing disk....remember that it was hard finding non-enterprise Seagate drives with warranties greater than 1 year then (a couple years ago ;)....got 2 year ones.... One drive went....after about a month we finally replaced it, ended up replacing all the drives as the seagates failed out. Luckily nothing was lost. Not like our 9990V where it got hit by a failure that (apparently) did corrupt some LUNs (only those attached to Windows.) Couple weeks ago had to swap LUNs on various Unix servers....there was a mix of vxvm, zfs and ASM. I did the first two, and the prep work (format, partition, chown device nodes to oracle) so that the DBAs could do the last one. The hard one involved suncluster. Something that never came around for me to get training, and everybody that knows it has quit. Just as the only person that knew our SAN was encouraged to quit, and now we're keeping parts of our old SAN as its too expensive to totally replace it all at once. Once things LUNs were swapped, Hitachi reformatted that portion of the array. There was some mumbling about how we had provisioned it, etc....but we had purchased it from Oracle, who then held us hostage because we had also purchased support from them. But, on support all we could get were drives replaced....no firmware updates, no upgrades, etc. I suspect it led to maximizing usable space from it over everything else. The fork lift upgrade of our entire datacenter was quoted at $11M, we're only getting $5M worth. Anyways....while I was looking trying to requisition for replacement drive(s).... WD Purple had recently appeared on the scene...and on 'paper' it seemed to be better than WD Red in all ways.... Purple is for arrays up to 8 drives, while Red is for arrays up to 5 drives (that was with NASware 2.0, 3.0 raises it to 8.) And, Purple is designed for always-on, low-power, 24x7 operation, in high-temperature environments...and supports TLER. (the Seagates we had didn't...had gotten the array to take a drive back for a few days a time, before I pushed to it replaced.) They are also write optimized (specifically for write-intensive, low bit rate, high stream count application....) Which I'm finding means they perform poorly on reads. What what drives I'll get when I get to upgrading my home storage.... Oh yeah...forgot the important reason....at the time, CDW(-G) had WD Purple drives priced lower than WD Red. (appears to still be that way....) Hmmm, guess its also that way for Amazon....for 2 TB drives. I upgraded a mirror to a pair of 3 TB WD Red's at home last month.... On 09/17/14 11:39, Ski Kacoroski wrote: > On 09/17/2014 06:27 AM, Matt Simmons wrote: >> With the response time discrepancy, I would think it's likely that the >> fast response time on writes is due to caching. >> >> You say that this is the only volume on the SAN seeing these times. I >> assume that this is the only host accessing that LUN? Is that the only >> LUN being accessed by this host? >> >> --Matt >> > > Yes, it is the only host using this lun. The host is accessing multiple other > luns (data, os, etc.). Going to talk with the vendor of the SAN today. I > will keep people posted what I find out. > > cheers, > > ski > -- Who: Lawrence K. Chen, P.Eng. - W0LKC - Sr. Unix Systems Administrator For: Enterprise Server Technologies (EST) -- & SafeZone Ally _______________________________________________ Discuss mailing list [email protected] https://lists.lopsa.org/cgi-bin/mailman/listinfo/discuss This list provided by the League of Professional System Administrators http://lopsa.org/
