From: robin_listas@es.invalid   
      
   On 2026-02-25 16:25, Paul wrote:   
   > On Wed, 2/25/2026 7:28 AM, Carlos E.R. wrote:   
   >> On 2026-02-23 05:11, Paul wrote:   
   >>> When you connect 3.5" drives in their own wall-powered enclosures,   
   >>> those don't draw hub current, so an unpowered hub works just as   
   >>> well as a powered hub for that case.   
   >>>   
   >>> That would be one reason for me recommending the "larger enclosures"   
   >>> -- they tend to be self-contained solutions. And if they have a   
   >>> fan ? You're laughing.   
   >>   
   >> They have their own troubles.   
   >>   
   >> I have two Yottamaster bay, 5 slots each, implementing an 8 disk software   
   raid 6 array (the extra slot is for the hot spare).   
   >>   
   >>   * The firmware sends the disks to sleep at 10 minutes.   
   >>   * A box can suddenly go away for an instant, destroying the array,   
   needing manual action to recover.   
   >>   * All the disks get almost the same name, depending on what type of name   
   you choose (Linux).   
   >>     It is difficult to make sure which hard disk of the lot one is   
   accessing.   
   >>   * Insert or remove one disk, there is a reset and name change of all the   
   disks.   
   >>   * It has a fan, but it is not enough. I had to piggyback an external fan.   
   >>   * It is slow.   
   >   
   > There could be a SATA mux in the thing. I don't   
   > know what happens when a SATA mux also has to   
   > support HotSwap.   
   >   
   > If you use separate, one-drive enclosures, chances are   
   > that could give a bit more predictable behavior. Then   
   > you need a USB3 hub "that won't go away" :-)   
      
   Absolutely, but I wanted to reduce the number of cables and power supplies. I   
   thought it would work better.   
      
   I'll write here the report I prepared for Amazon. I used DeepL to translate.   
      
      
   +++-----------------------   
   Yottamaster   
      
   Model FS5 according to the label. But according to the manual, it is   
   ‘Y-Focus Series 5-Bay’ FS5U3.   
      
   Mechanically, the case is very good, although the discs are somewhat difficult   
   to remove as there is no good handle.   
      
   The firmware leaves much to be desired. The chipset is idVendor=152d,   
   idProduct=0578 (JMicron Technology JMS578 SATA 6Gb/s). It can be flashed, but   
   someone has to find the right package for this enclosure.   
      
   The enclosure shuts down the disks after 10 minutes, even if you are running   
   the long SMART test at that moment. It aborts it. To run the long test, you   
   have to write a script that launches an operation on the disk every 8 or 9   
   minutes.   
      
   The enclosure replaces the disk's serial number with its own (not always),   
   making it difficult to identify the disks from the operating system (Linux in   
   my case). There is SMART support, but some commands have to be run twice, or   
   it will return the    
   previous result. With smartctl, you can find out the real serial number:   
      
      
   	smartctl -a /dev/disk/by-path/‘pci-0000:00:14.0-usbv3...’ | grep   
   ‘Serial Number:’   
   	smartctl -a /dev/disk/by-path/‘pci-0000:00:14.0-usbv3...’ | grep   
   ‘Serial Number:’   
   And also with:   
      
        hdparm -I /dev/disk/by-path/‘pci-0000:00:14.0-usbv3...’ | grep   
   ‘Serial Number’   
      
   You can hot-swap disks, but this causes a reset that makes the operating   
   system scan the disks again, and they appear in the system with random   
   identifiers. A disk you were using may appear with a different ID, and the   
   mount will break. You have to mount    
   partitions by UUID or another identifier that is recorded on the disk.   
      
   Having two boxes like I do makes the mess even bigger. In Linux, you have to   
   use ‘/dev/disk/by-path/’ because it is the only one that includes all the   
   disks. In addition, the identifiers   
      
        /dev/disk/by-path/‘pci-0000:00:14.0-usbv3-0:1:1.0-scsi-0:0:0:0’   
      
   are maintained between boots and refer (apparently) to the same serial number;   
   they are stable. I understand that they refer to the hole in the box where   
   each disk is located and the USB port to which it is connected on the   
   computer. But they change when    
   a disk is removed.   
      
      
   To find a disk in the box, use:   
      
        while :; do smartctl -a $THEDISK; done 1>/dev/null   
   or   
        while :; do smartctl -a $THEDISK; delay 2; done 1>/dev/null   
      
   This causes the LED to flash slightly differently than during normal disk   
   activity, but it is better if all disks are inactive.   
      
   You can see a discussion of the topic in ‘nntp:alt.os.linux’ subject: *   
   Error sending ATA command IDENTIFY DEVICE   
      
   I have purchased two boxes to make a software RAID, and there is a problem:   
   from time to time one of the boxes disappears for a moment, and the operating   
   system rescans the disks. This causes the RAID to disconnect the four disks   
   and you have to reinsert    
   them manually. This is the beginning:   
      
   May 06 09:47:41 Isengard kernel: usb 2-1: USB disconnect, device number 2   
   May 06 09:47:41 Isengard kernel: sd 2:0:0:0: [sdh] Synchronising SCSI cache   
   May 06 09:47:41 Isengard kernel: sd 2:0:0:0: [sdh] Synchronise Cache(10)   
   failed: Result: hostbyte=DID_NO_CONNECT driverbyte=DRIVER_OK   
   You can find the reported problem on the internet; for example, on Reddit:   
   ‘[HELP] External 4bay yottamaster keeps disconnecting’   
      
   The measured speed (hdparm -tT) of the array is 179.88 MB/sec.   
      
   -----------------------++-   
   >   
   > *******   
   ...   
      
   ...   
      
   ...   
      
   --   
   Cheers, Carlos.   
   ES🇪🇸, EU🇪🇺;   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)