Please consider a donation to the Higher Intellect project. See https://preterhuman.net/donate.php or the Donate to Higher Intellect page for more info. |
Difference between revisions of "Seagate ST43400N issues on SGI"
(Created page with "Is anyone else running one of these disks on one of their SGI machines? We are on our third one now (the first two having been swapped out) and each of them has developed err...") |
|||
Line 26: | Line 26: | ||
[[Category:SGI]] | [[Category:SGI]] | ||
+ | [[Category:Hard Disks]] |
Revision as of 15:44, 27 July 2019
Is anyone else running one of these disks on one of their SGI machines? We are on our third one now (the first two having been swapped out) and each of them has developed errors like this:
Nov 18 02:20:03 2A:iris unix: dks0d2s4 (/local): retrying request Nov 18 02:20:04 2A:iris unix: dks0d2s4 (/local): Media error: Unrecovered data block read error (asc=0x11, asq=0x0), (data byte 28), Block #-3125472 (666768) Nov 18 02:20:04 2A:iris unix: dks0d2s4 (/local): retries exhausted Nov 18 02:20:04 2A:iris unix: dks0d2s4 (/local): Recovered Error: Read data recovered with retries (asc=0x17, asq=0x0), (data byte 21), Block #-3125472 (666768)
after a short time in use. We are running IRIX 5.2 on a 4D/380 VGX machine. Is there anything I might have overlooked that might cause this?
We have six Crimsons with 2 of these 2.9gb drives each. All of them give these errors. Also, all of them eventually fail to boot with "Can't read block 1; can't find unix" errors; if you fsck the heck out of root, it will then boot OK until the next time you take it down. When we return the drive to the local vendor, she/he swears that the drive looks normal in a non-SGI machine. The upshot is that even tho it's swell to have 2.9 gig in one (or two) spot(s), it's not worth it in terms of rework, reliability, etc. We're recommending that the buyers go back to the lower-capacity drives which almost never get errors, and hoping that the Seagate 4gb drives get to be reliable enough to use for critical stuff... I spent 4 hours wrestling with one of the darned things today on what should have been a simple 1-hour troubleshooting job.