We have a 2006 Sun X4500 with Hitachi 500G disk drives. Its been running for over four years and just now fmadm & zpool reports a disk has failed. No data was lost (RAIDZ2 + hot spares worked as expected.) But, the server is out of warranty and we have no hardware support on it. I found the Sun part numbers for X4500 replacement drives here: http://docs.sun.com/source/819-4359-18/CH3-maint.html#50499627_22785 I''ve been waiting for a quote on one for a while. In the meantime, I figure it would be faster and cheaper just to buy a disk drive directly. The /usr/bin/hd program (from the SUNWhd package) gives this info for the failed disk: c5t0d0p0 VN67ZAKLT1MH ATA HITACHI HDS7250S AJ0A 25 C (77 F) It''s a Hitachi Deskstar 7K500 model. But, that particular model doesn''t seem to be around anymore (P7K500 is all I could find.) I notice the Deskstar lines are the cheapest, "consumer" level drives--I wonder why they didn''t spec the Ultrastar server/nearline drives instead? I read about issues with zfs and write cache: http://opensolaris.org/jive/thread.jspa?messageID=449656&tstart=0 Does anyone have any bad experiences replacing a disk on an X4500 with a non-Sun Hitachi? The hdadm tool reports the write cache is enabled on all the disks. Are their any customized firmware on the Sun disks that make them safer for using write cache? Thanks -- This message posted from opensolaris.org
----- "Doug" <dy7t at yahoo.com> skrev:> Does anyone have any bad experiences replacing a disk on an X4500 with > a non-Sun Hitachi? The hdadm tool reports the write cache is enabled > on all the disks. Are their any customized firmware on the Sun disks > that make them safer for using write cache?Compared to what you have now, I gues anything will do. It doesn''t need to be 500 gigs, just use something >= that size, preferably larger, in case the 500 gigs turns out to be 499. Best regards roy -- Roy Sigurd Karlsbakk (+47) 97542685 roy at karlsbakk.net http://blogg.karlsbakk.net/ -- I all pedagogikk er det essensielt at pensum presenteres intelligibelt. Det er et element?rt imperativ for alle pedagoger ? unng? eksessiv anvendelse av idiomer med fremmed opprinnelse. I de fleste tilfeller eksisterer adekvate og relevante synonymer p? norsk.
On Wed, May 12, 2010 at 09:34:28AM -0700, Doug wrote:> We have a 2006 Sun X4500 with Hitachi 500G disk drives. Its been running for over four years and just now fmadm & zpool reports a disk has failed. No data was lost (RAIDZ2 + hot spares worked as expected.) But, the server is out of warranty and we have no hardware support on it.Well - had the same thing here (X4500, Q1 2007) 2-3 times couple of month ago. The ''too many errors'' msg ringed some bells: do you remember the race condition problems in the marvell driver (IIRC especially late u3, u4) which caused many ''bad ...'' errors in the logs? So I simply checked the drive in question (Q&D 2xdd over the whole disk and checked, whether an error occured). Since not a single error or bad performance I put it back and no wonder, it is still working ;-) ). Your situation might be different, but checking may not hurt - your disks might be a victim of an SW aka ZFS error counter... Have fun, jel. -- Otto-von-Guericke University http://www.cs.uni-magdeburg.de/ Department of Computer Science Geb. 29 R 027, Universitaetsplatz 2 39106 Magdeburg, Germany Tel: +49 391 67 12768
On 05/13/10 08:55 AM, Jens Elkner wrote:> On Wed, May 12, 2010 at 09:34:28AM -0700, Doug wrote: > >> We have a 2006 Sun X4500 with Hitachi 500G disk drives. Its been running for over four years and just now fmadm& zpool reports a disk has failed. No data was lost (RAIDZ2 + hot spares worked as expected.) But, the server is out of warranty and we have no hardware support on it. >> > Well - had the same thing here (X4500, Q1 2007) 2-3 times couple of > month ago. The ''too many errors'' msg ringed some bells: do you remember > the race condition problems in the marvell driver (IIRC especially late > u3, u4) which caused many ''bad ...'' errors in the logs? So I simply > checked the drive in question (Q&D 2xdd over the whole disk and checked, > whether an error occured). Since not a single error or bad performance > I put it back and no wonder, it is still working ;-) ). > >This backs up my experiences with x4500s. I have had several drives "fail" which I have taken off line and thrashed with format for a couple of days without finding any errors. Out of 9 or 10 "failures" only one was FUBAR. -- Ian.