chickensoup Posted July 7, 2012 Share Posted July 7, 2012 Renamed topic from Mover spun up all drives, nothing to be moved? - see update 9 July 2012. I just happened to be up late tonight and noticed my mover spun up every single drive in the array even though the cache drive was empty, I also happened to have putty open and it doesn't appear anything was moved. According to the FAQ "If the mover finds no files to move, and the disks are spun-down, the disks will not spin up." The size of my cache drive on the WebUI appears to show something on the disk, though a quick ls -a confirms that it should be empty. Cache drive on WebUI, Total Size: 146,522,292 Free: 146,484,980 (Difference of 37,312). This is after the mover had been run, I didn't notice the size prior though I am fairly confident the drive was empty & that is what the log shows- 16 seconds of moving.. nothing. I do have a Gigabyte motherboard and have not checked if the BIOS is set to backup to HDD- could this be what is "hiding" on my cache drive? Even if it is, it shouldn't affect the mover? All my drives were still spinning at 3:37am and thought it was a little weird as I was only watching one documentary, lsof showed it was the only open file so I did a manual spin down of all drives (not realising the mover was about to initiate anyway). 3 Minutes later the mover spun up all the drives but moved nothing..? I was also curious about the way "sdh" was being spun down (Tower emhttp: shcmd (55): /usr/sbin/hdparm -y /dev/sdh >/dev/null) and it turns out this is my cache drive- is this normal? Update - 9 July 2012 OK it seems I have another (possibly related problem). The situation is this: No drives were spinning at all (as per the WebUI). I opened one of my shares ("Storage") which is only supposed to be disk7 WebUI shows all disks spun up (disk1 - disk7) and parity drive. Cache drive did not. "Storage" is setup as follows: included disks (disk7), excluded disks (disk1,disk2,disk3,disk4,disk5,disk6), use cache (yes), allocation (high-water). According to the webUI and Windows disk7 has ~250GB free. Why would all the drives spin up in this scenario? The folder "Storage" only appears on disk7, it is not shown on any other drive. Included updated syslog 9-7-12, shows nothing new other than telnet login. syslog9-7-12.txt Link to comment
dgaschk Posted July 7, 2012 Share Posted July 7, 2012 Edit: I was also curious about the way "sdh" was being spun down (Tower emhttp: shcmd (55): /usr/sbin/hdparm -y /dev/sdh >/dev/null) and it turns out this is my cache drive- is this normal? Yes. Link to comment
Joe L. Posted July 7, 2012 Share Posted July 7, 2012 The statement in the wiki depends on the specific version of unRAID you are running. It will spin them up if it thinks there is something to move. It did spin up the drives on earlier versions of unRAID. Link to comment
chickensoup Posted July 8, 2012 Author Share Posted July 8, 2012 The statement in the wiki depends on the specific version of unRAID you are running. It will spin them up if it thinks there is something to move. It did spin up the drives on earlier versions of unRAID. When you say "It will spin them up if it thinks there is something to move." does this mean if the cache drive is anything other than completely empty, the mover will spin up all drives? Because it doesn't appear that anything was moved, nor does it appear that the drive is completely empty (in size) though there is no files/folders on the cache drive at all. Link to comment
chickensoup Posted July 8, 2012 Author Share Posted July 8, 2012 Renamed topic & updated OP as I seem to have other problems. Didn't bother to make a new topic as it is likely related. Link to comment
RobJ Posted July 9, 2012 Share Posted July 9, 2012 I don't see anything wrong in your syslog. You do have a Gigabyte HPA on your 150GB Cache drive, but that does not matter. There have been bugs fixed in the recent release candidates of the v5 betas, in connection with incorrect spin ups and downs. I have no idea whether they are involved with what you are seeing. Your Spinup groups look very unusual, that is all I can tell from the syslog. Check the Spinup group label you have assigned to each drive and make sure they are all unique. Default labels now are 'host0' for the parity drive, 'host1' for Disk 1, 'host2' for Disk 2, 'host3' for Disk 3, etc. If you find any that are equal, then they will cause each of the grouped drives (equivalent labels) to spin up, if any one of them spins up. Link to comment
chickensoup Posted July 9, 2012 Author Share Posted July 9, 2012 Ahh, I think I know what the problem might be.. perhaps I have misunderstood the way the spinip groups are supposed to behave. It was my understanding that spinup groups were created to minimise load on the PSU, for instance when you initiate a parity check or manually tell all drives to spin up via the webui. Since I only have 8 data drives and a fairly decent power supply I set all of my drives to the same spinup group in the thought that it would minimise the wait time when accessing a share that required multiple drives. I didn't realise these drives had to spin up together if in the same group. I'll check the settings when I get home but I think is is what is causing the issue. Link to comment
chickensoup Posted July 10, 2012 Author Share Posted July 10, 2012 I disabled spin-up groups and the problem has gone away. Obviously it is my settings, will look more into it another day. Renamed topic to solved, will start a new topic if my mover issue is unrelated. Thanks guys Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.