DZMM Posted December 27, 2017 Share Posted December 27, 2017 One of my 5TB drives is showing signs of failing - reallocated sectors, read errors, offline uncorrectable etc creeping up. I want to pull it and try pre-clearing again to see if this will fix - or is this wasted effort? If I RMA, I will be without a drive for a week or so. What's the best practice for using my parity drive to replace the drive in the array until the replacement drive arrives, or is the parity drive only for rebuilding the array when I get the new drive? Thanks Edit: Or, is it easier to move the data off the 5TB drive onto the other disks and temporarily shrink the array by 5TB until the new disk arrives? Link to comment
pwm Posted December 27, 2017 Share Posted December 27, 2017 If the numbers are increasing and you still have warranty, then you don't want to try to repair and keep the drive. You shouldn't play with the parity drive. It can be used to emulate a missing drive, but at the cost of lost redundancy. If you only have one parity drive, and the parity drive fails then you lose the ability to emulate the missing drive while waiting for the replacement. And if any of the other data drives fails, then you lose the content of that drive AND the ability to emulate the missing drive - so you potentially lose the content from two data drives. If you can, then I would recommend you move the data off the drive and shrink with one drive. Link to comment
DZMM Posted December 27, 2017 Author Share Posted December 27, 2017 Thanks - I think the shrink array option is the best one, the more I think of it. I've got 4.7TB of spare space and 4.9TB of files on the drive, so I'll move some none critical files onto my wife's laptop to free up the missing space. Link to comment
Frank1940 Posted December 27, 2017 Share Posted December 27, 2017 A few years back, I had a similar situation. I just turned the server off, pulled the drive and waited until I got the return. My experience has indicates that, apparently, the drive manufacturers do not examine/analyze the drives on receipt. The reason being that they usually ship the replacement on the same as the arrival of the return. They must be assuming that most folks don't pull a "working" drive without a good reason for thinking that it is defective. However, I would think that they do track all customers to make sure that some aren't 'abusing' (or "working") the warranty system and if you make that list, it might take longer to get your replacement. Link to comment
DZMM Posted December 27, 2017 Author Share Posted December 27, 2017 Turning off isn't an option - the server controls our internet (pfsense VM), TV (tvheadend), smarthome (ha) and more! Link to comment
TUMS Posted December 27, 2017 Share Posted December 27, 2017 If it were me I would just buy a new drive. Then worry about replacement after I had the new one safely running. Link to comment
DZMM Posted December 27, 2017 Author Share Posted December 27, 2017 10 minutes ago, TUMS said: If it were me I would just buy a new drive. Then worry about replacement after I had the new one safely running. I would if I could - I've used up all my SATA ports, so I'll just end up with an unused drive Link to comment
JonathanM Posted December 28, 2017 Share Posted December 28, 2017 6 minutes ago, DZMM said: I would if I could - I've used up all my SATA ports, so I'll just end up with an unused drive You've backed yourself into kind of a bad corner. I'd correct this ASAP, either by higher capacity drives, or if you have the physical space, adding a controller. It's not a good idea to be running without available extra slots to maneuver around when you get into a jam with recovery or prepping a new drive. Does your warranty allow for advanced replacement instead of waiting? You say it's not acceptable to down your server for any length of time, but you haven't put yourself in a situation where high availability is easy to achieve. Warranty replacement drives are a crapshoot, you really need to test it thoroughly before trusting it to safeguard the rest of your data. Keep in mind if you have an unexpected failure while running degraded, you will lose data. If your replacement drive isn't perfect, and one of your other drives decides to die, you will lose both. If I were you, I'd put a new 8TB in the array, and use the warranty replacement 5TB for backup of critical data. Link to comment
Frank1940 Posted December 28, 2017 Share Posted December 28, 2017 3 hours ago, jonathanm said: If I were you, I'd put a new 8TB in the array, and use the warranty replacement 5TB for backup of critical data. +1 (Or you could save it to replace that 2TB drive as you run out of space or a replacement if one of the other two 5Tb drives fail.) Link to comment
DZMM Posted December 28, 2017 Author Share Posted December 28, 2017 4 hours ago, Frank1940 said: +1 (Or you could save it to replace that 2TB drive as you run out of space or a replacement if one of the other two 5Tb drives fail.) You're right.... I'm going to bring forward my plan for next year: 1. remove my two 250GB SSD cache drives and replace with 1x500GB to free up a slot 2. replace the 2TB with a 8TB Link to comment
BobPhoenix Posted December 29, 2017 Share Posted December 29, 2017 I always keep cold spares that have been precleared available. Then I can change out a bad drive instantly and order a new or RMA the bad as a replacement cold spare. Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.