G+_Shawn Ashe Posted October 28, 2017 Share Posted October 28, 2017 Hey, I've got a Synology DS1513+ with 4 WD Red 3TB drives in it (one slot open). When one drive went bad last week, I ordered two immediately. I replaced and repaired the bad drive without incident (yay- DSM!) Now.. the question I have is what do I do with the 5th drive? Store it on the shelf for the next failure? Put it in the 5th slot as a hot spare? Add it to the NAS (SHR raid, currently one disk fault tolerance) This unit is usually used as a remote backup (and video server) at a vacation home, so I don't usually have immediate access when failures occur. I was able to log in remotely to shut it down when the failure occurred, but now I'm trying to decide the best use of the spare disk. Does it cause damage to a disk to be sitting powered up as a hot spare compared to sitting on a shelf? Is a hot spare of any value as compared to just adding it to the raid set? What are your ideas/ Link to comment Share on other sites More sharing options...
G+_Travis Hershberger Posted October 28, 2017 Share Posted October 28, 2017 1 disk fault tolerance with 3TB drives is a bad idea, I'd put in that 5th drive and bump it up to 2 disk fault tolerance. Adding the extra drive as a hot spare is just going to automate failure. Link to comment Share on other sites More sharing options...
G+_Golden Retriever Posted October 28, 2017 Share Posted October 28, 2017 Travis Hershberger run GRC spinrite on it whatever your final choice become Link to comment Share on other sites More sharing options...
G+_Ben Reese Posted October 28, 2017 Share Posted October 28, 2017 I'd vote for the 2-disk redundancy too. Hotspair gets the 2nd vote. Of the bad drive is under warranty still or even close, I'd contact WD. They replaced a failing 2 TB drive I had that was a few months out of warranty. Even sent me the replacement so I could move my data before sending the drive back. Link to comment Share on other sites More sharing options...
G+_Shawn Ashe Posted October 29, 2017 Author Share Posted October 29, 2017 Thanks, I think I agree with adding it to the volume. Link to comment Share on other sites More sharing options...
G+_Jim Lloyd (AudioMaveric Posted October 31, 2017 Share Posted October 31, 2017 Hot Spare! Let the drive auto-rebuild for you. I have had many RAID array experiences where drives seem to fail when I am out of the office... vacation, weekend, sleeping. Link to comment Share on other sites More sharing options...
G+_Travis Hershberger Posted October 31, 2017 Share Posted October 31, 2017 Jim Lloyd you've gotten very lucky if these failures have happened with drives of 500GB or larger. You're most likely automating failure at that point. smbitjournal.com - Hot Spare or a Hot Mess Link to comment Share on other sites More sharing options...
G+_Jim Lloyd (AudioMaveric Posted October 31, 2017 Share Posted October 31, 2017 Travis Hershberger Interesting article. To each, I guess... . I've been in the IT business a long time. Drives fail for various reasons. RAID can drop and blow your data, regardless. It depends on the maturity of the technology and level of investment, at the time. So, anyone storing data in any fashion assumes a level of permanent data loss risk. That being said, drives fail. I'd rather have a drive waiting to insert itself and rebuild, ASAP. . To re-answer Shawn's question, "I would keep the drive installed as a 'hot spare', 'warm spare', whatever you like to call it. Let it start rebuilding, as soon as possible. Replace the bad drive, as soon as possible. . To answer your inference, pull out your checkbook. Set up two RAID6 servers with two standby spares, mirror them, and put them in separate countries. Maybe, your data will be safe... (:>) Link to comment Share on other sites More sharing options...
G+_Ben Reese Posted October 31, 2017 Share Posted October 31, 2017 Jim Lloyd? wouldn't it be better to add it as another redundant drive that to have it as a hot spare? Basically, would it be better to have 4 drives in RAID 5 with a hot spare or 5 drives in RAID 6? I'm a DBA, not a storage admin, so I'm legitimately curious. My thought is that reducing the risk of failure is better than automating for it. Link to comment Share on other sites More sharing options...
G+_Jim Lloyd (AudioMaveric Posted October 31, 2017 Share Posted October 31, 2017 Ben Reese That is the big question... RAID6 buys one the luxury of losing two consecutive drives, before a drive replacement, without data loss. A failed drive still needs to be replaced and rebuilt with the missing data. Which scenario best fits your environ ment? Link to comment Share on other sites More sharing options...
G+_Ben Reese Posted October 31, 2017 Share Posted October 31, 2017 Jim Lloyd right, that's what I'm asking. When does it make more sense to use a hot spare than have another drive for redundancy? Again, not an expert, but it seems a hot spare makes most sense if... You have SSDs that could rebuild the array quickly but you're limited to the number of writes? Have multiple storage groups where the hot spare could play backup for everyone? You have an older drive that you don't trust for daily operations but hope that it could help out in a failure? Anything I missed? In this case, I would think it would be better to convert from SHR-1 to SHR-2 than to have the hot spare. UNLESS you plan on switching back to SHR-1 later, which I don't think is possible. (opinion based on logic, not experience.) Link to comment Share on other sites More sharing options...
G+_Jim Lloyd (AudioMaveric Posted October 31, 2017 Share Posted October 31, 2017 Ben Reese , you are thinking in the correct direction. Most RAID controllers now use the spare to fill in for the dead drive, and then go back to spare when the failed drive is replaced and finishes rebuilding. I don't know of a scenario where one drive can be assigned to multiple arrays, though. I chose a 6-drive RAID5 with two spares on my DS1815+ config. That works out about the same, I think... old school. . Back 10+ years ago, it was not uncommon for IT to schedule drive replacements in an array, sort of like replacing UPS batteries on a schedule. One at a time, until they rebuilt. Sometimes took more than a work week. . I have seen many ideas on how to approach maintaining an array, but nothing beats backing one up to another location! Even Padre uses two Synologies in a mirror. I had a RAID controller (20 years ago) fail writing garbage to all the drives as it did. There was nothing left to salvage. Planning seems to have the "N+1" that surprises you. Plan as best you can justify. There is the 'perfect' solution, and the one you can afford. Link to comment Share on other sites More sharing options...
G+_Shawn Ashe Posted November 2, 2017 Author Share Posted November 2, 2017 Power was out all week here, so did not move forward yet. Part of my thoughts had to do with degradation of the drive while its not being used. Sitting on a shelf un-powered, I'm sure it will age (caps dry out) if not used. Sitting as a hot/warm spare, does that age the drive just as if its being used in the raid? Is it being exercised by DSM while sitting there? Link to comment Share on other sites More sharing options...
G+_Shawn Ashe Posted November 4, 2017 Author Share Posted November 4, 2017 Well.. looks like the only way to change from SHR 1 to SHR 2 disk fault tolerance is to completely rebuild. So its probably best to set this as a hot spare? Link to comment Share on other sites More sharing options...
G+_Ben Reese Posted November 4, 2017 Share Posted November 4, 2017 Completely rebuild and in rebuild the RAID, or rebuild the disk group/volumes and lose all the data? Either way, I guess that would have to be your judgment call. Link to comment Share on other sites More sharing options...
G+_Travis Hershberger Posted November 4, 2017 Share Posted November 4, 2017 Shawn Ashe Adding a hotspare is automating failure. You don't want a rebuild happening with only a single parity drive unless you've already take a backup! Link to comment Share on other sites More sharing options...
Recommended Posts