February 24, 201511 yr As far as what fixed this, it could very well be the change we made in 14a to poll_spindown to not tie that event to checking disk attributes (SMART). If this will be the case, then I will nostalgically think back to my post sometime beginning of Jan when I first suspected this to be the issue...
February 25, 201511 yr Looking at the announcement thread of b14a I think the good news is that now we know the root cause of this will be something with 'poll_spindown' event. This also makes sense historically: 'poll_spindown' event has been added between b11 and b12 and the issue popped up in b12. What actually scrares me a bit is that Tom was not even mentioning this issue in the b14a announcment (I mean that there is a potential fix for this) which makes me feel the issue was/is not really on his radar so this was/is not taken seriously. Now it probably will because with 14a it has got proven that something is wrong.
February 25, 201511 yr The issue isn't solved in 14a, it's just hidden. Tom made a post mentioning that the webGui isn't displaying the correct spin status right now in 14a. We're working on a 14b which may correct that and be the fix for this issue altogether. As far as this being on Tom's radar, it definitely has been and we've discussed it at length internally. The problem has been and always has been consistently reproducing this in a lab. Even yesterday after all my testing and going back and forth between different versions, I thought I had it all solved and then POOF, it starts happening again to me on 14a after the release. When an issue is reported per the guidelines posted here, we can usually find it and fix it fairly quick because if posting under those guidelines, it will contain the exact steps to reproduce it. When posts do not follow those guidelines or when the steps to reproduce don't work on our test systems, that's when issues like this can take longer to resolve and don't get mentioned in updates. What I or Tom say in a reply on a thread when trying to troubleshoot and what should get stated in an official announcement post are two different things. Hopefully this clarifies.
February 25, 201511 yr I hope you don't blame us not being able to report against those guidelines, because I think you are in the same shoes. We were/ are all trying to find a pattern and a clear way to reproduce, but I think by now we all know that the characteristic of the issue just doesn't allow this. Also, I think the guys here were actually putting huge efforts in trying to locate and catch the issue. You were also able to reproduce finally. So all what I was just trying to articulate that after all the time, effort, willingness and testing made here, it would be good feeling to see the issue acknowledged and recognized in the announcment threads as well, especially when there is a potential for being resolved.
February 26, 201511 yr I hope you don't blame us not being able to report against those guidelines, because I think you are in the same shoes. We were/ are all trying to find a pattern and a clear way to reproduce, but I think by now we all know that the characteristic of the issue just doesn't allow this. Also, I think the guys here were actually putting huge efforts in trying to locate and catch the issue. You were also able to reproduce finally. So all what I was just trying to articulate that after all the time, effort, willingness and testing made here, it would be good feeling to see the issue acknowledged and recognized in the announcment threads as well, especially when there is a potential for being resolved. Why would you think I blame you guys? Please take my post for face value. There was no "slam" to the community on that at all. The simple fact is without the guidelines being followed (or even when they are, if we can't recreate in our own test environments), the issue will take longer to track down and resolve. Without testers like you guys, this stuff would take even longer and more bugs might slip through without the reporting. We definitely appreciate it! As far as acknowledging it in the announcement thread, if you notice, beta 14b is now available and the first note is regarding the fix for this issue.
February 26, 201511 yr Beta14b has fixed my spin down issues. I know this problem was difficult to find and I appreciate it. Thanks. Gary
February 26, 201511 yr Beta14b has fixed my spin down issues. I know this problem was difficult to find and I appreciate it. Thanks. Gary THANK YOU FOR CONFIRMING! That's great to hear. Hoping to see more people chime in so we can officially mark this thread as Solved.
February 27, 201511 yr Beta14b has fixed my spin down issues. I know this problem was difficult to find and I appreciate it. Thanks. Gary THANK YOU FOR CONFIRMING! That's great to hear. Hoping to see more people chime in so we can officially mark this thread as Solved. I'm on 14b, have 5 WD drives and the spin down issue is resolved for me. I did use idle3ctl when that was posted and that seemed to help then though. But was under the impression the gui wasn't reporting correctly for one of those betas. So I'm not sure if idle3ctrl really helped.
February 27, 201511 yr Beta14b has fixed my spin down issues. I know this problem was difficult to find and I appreciate it. Thanks. Gary THANK YOU FOR CONFIRMING! That's great to hear. Hoping to see more people chime in so we can officially mark this thread as Solved. I'm on 14b, have 5 WD drives and the spin down issue is resolved for me. I did use idle3ctl when that was posted and that seemed to help then though. But was under the impression the gui wasn't reporting correctly for one of those betas. So I'm not sure if idle3ctrl really helped. Woo hoo! Thanks Derek for confirming!!!
February 27, 201511 yr Appears resolved for me as well in 14b. I am not coming home to disks spun up. Real test will be if I have disk up in the morning after watching shows/mover runs. But looks promising with others confirming. Thanks to all who helped snipe this one!
February 27, 201511 yr I'm also seeing drives shown as spun down when I expect them to be. Presuming that the state is being reported correctly (and access delays would suggest that this is so), then all would seem to be well. Only oddity - I have a drive which has been spun down for more than twelve hours, yet the drive temperature is still being reported. I know that unMenu made a special case for WD drives, which don't have to be spun up in order to read the temperature, but I don't believe that unRAID/dynamix does this ... so where is this temperature coming from?
February 27, 201511 yr I'm also seeing drives shown as spun down when I expect them to be. Presuming that the state is being reported correctly (and access delays would suggest that this is so), then all would seem to be well. Only oddity - I have a drive which has been spun down for more than twelve hours, yet the drive temperature is still being reported. I know that unMenu made a special case for WD drives, which don't have to be spun up in order to read the temperature, but I don't believe that unRAID/dynamix does this ... so where is this temperature coming from? When I could not tell for sure whether a drive was spun-up or not, I watched the temperature. If it is significant higher than ambient and/or changes (more than a change in the amibent would account for), you can be almost certain that it is spun-up. Hope this helps...
February 27, 201511 yr Hi, issue not resolved here, some drive seems to be spin down but they are still reporting the temperature, and my sever woon't enter into S3 automatically. If needed I can provide my logs.
February 27, 201511 yr Hi, issue not resolved here, some drive seems to be spin down but they are still reporting the temperature, and my sever woon't enter into S3 automatically. If needed I can provide my logs. S3 is not supported by us just yet and not sure how that plugin may affect this. The temp reporting issue is fixed in next release. Your drives should spin down fine. Please remove plugins and test again for spin down.
February 27, 201511 yr I have one disk left that still refuses to automatically spin down. Its a Toshiba 5TB.
February 27, 201511 yr Haven't seen this happen on beta12 or higher, :-). Confirming solved for me as well. Feb 27 03:41:57 Tower logger: mover finished Feb 27 03:56:38 Tower kernel: mdcmd (68): spindown 2 Feb 27 03:59:00 Tower kernel: mdcmd (69): spindown 0 Feb 27 03:59:01 Tower kernel: mdcmd (70): spindown 1
February 28, 201511 yr I can confirm that the issue is still present for me. All HDD seems to be spundown, but when I check the power consumption with my ups, it's too high so I've tried to access some files, witch should introduce a lag (due to the spinup process) but I could read my files without any delay.
March 3, 201511 yr Author This doesn't seem resolved for me. I *see* my drives still spun up in the gui.
Archived
This topic is now archived and is closed to further replies.