i was used to using hdparm to spin down my raid after a certain of inactivity. is that still relevant on nethserver / safe to do it this way on NS?
/sbin/hdparm -S 255 /dev/sdb /dev/sdc /dev/sdd
man
Put the drive into idle (low-power) mode, and also set the standby (spindown) timeout for the drive
NethServer is a bunch of packages/tools for writing configurations in a controlled way for a lot of different softwares. Therefore, if you have experienced nice results with CentOS 7 without compromising performance, when you’ll have the reverse setup command feel free to try.
However… I don’t know how much’s the size of disk, and disk type (rotary? SATA? 2,5 or 3,5"?. Also, i don’t know if you’re using (if a server) a md array or if your setup is more similar to a router or… what else?
If power consumption is your drive, SSD are quite “greedy” during data transfer (max power drain 3-6w, depending by size of disk, source, data referred to 2,5" SATA)) compared to HDD which strolls out from 1.4 to 1.7w (source, data referred to 2,5" SATA). On the other side, during idle or sleep energy drain is far more interesting for SSDs: 0,002 in sleep (compared to 0.1), increasing to 0.055 to 0.110 for idle (against 0.5) and with an “typical active” drain of 0.07 to 0.150, compared to 1.5-1.7 of HDD. So, most of the time, SSDs gains some energy savings that can come handy.
Also, switching to SSD add a nice plus about hit and vibration resistance, usually higher.
Ever rose has it’s torn… SSDs are “timed”, the number of writings is defined and also with a large enough cache, a power interruption may increase chances to create inconsistent state of the filesystem.
hey, thanks for that detailed explanation! did not cross my mind that one also could put sdds to sleep. guess that wont make any sense with my hardware tough.
system/os is on a 2.5 ssd 60gb
raid 5 = 3 x Seagate NAS HDD 3TB, SATA 6Gb/s 5900rpm
still got a WD GREEN 240GB in there but think i will take it out
i am using this pc for gateway router / wlan and web hosting (web stuff is only on the ssd)
the files on the raid are only accessed per nfs, also i am the only one on this machine.
not 33 people reading/writing all the time.
i was happy with those hdparm settings on the old centos machine. applied it now to NS. how would i go about making this persistent that it survives a reboot?
AFAIK no, that kind of parameters should not be affected from the configuration engine brought up from NethServer, but don’t take my knowledge as a reliable source of information. Hoping that someone about dev team could provide more valuable info.
Also, take a run into github 
you may look for hdparm into NethServer project.
will do - thanks
Now you’re right in the center of the debate: Do drives die quicker if they are left spinning continuously, or does frequent stop/start cycles do more damage.
Good luck finding any consensus. 
Cheers.
eh 
consensus is hard to find these days - humanity would be better off sticking to data and facts
FWIW my heart is on spinning continuously: IMVHO it’s the more effective way to come closer or overtake by far the MTBF of hard drives. Correct kind of disks (desktop ones die always too soon), UPSes and nice temperature are important contributions to reach “better never than late” goal for hardware failure.