Page 1 of 1

Drives pulsing every second or so like a heartbeat

Posted: 17 Aug 2023, 03:01
by d3x0r
I've had an F4-210 for a while; I had 2 drives in it, raided, and when those got filled, added a couple more drives.
When I initially setup the new drives with a partition, I wasn't using them, nothing had a map to the new share even, and the drives were pulsing activity, almost like a heartbeat (trying to think of how to represent the sound... du-du... du-du... du-du ... ) continuously, even after a reboot. I eventually connected a share and wrote some data to the drive, and the problem stopped.

As a side issue - I had updated the firmware, but the web interface had stopped working ( viewtopic.php?f=76&t=4655 ), didn't really need to make any changes and everything else was working so I just continued.

I was recently transferring a bunch of files from a couple computers to the new drives' shares, and crashed the NAS; it locked up and wouldn't respond to anything, after rebooting, the second partition went away. (I called it appdata2). I got the recent firmware update ( https://support.terra-master.com/downlo ... uct=F4-210 ) and from the .bz2 versions, managed to extract the www interface, and update the drive with it... after fiddling with it for a while, managed to find the raid, run an fsck on it, and get it mounted... and then after updating the permissions, got the drive all back up. I then did a install of the latest firmware ( I had installed the web interface one version back actually), and after rebooting, all worked well; got the files I wanted to transfer transferred.

But, it was noticed this morning that the pulsing behavior is back. There aren't any tools I can find that can tell me what is causing the activity, although I did find that `cat /sys/class/block/md1/stats` the the write counters are ticking up a couple ticks every second, so I can assert that there are spurious writes happening to the drive, but I don't know why, what, where... though I can say it's the latest ARM firmware.

if I run top

Code: Select all

 1843     1 root     S     513m  53%   1% syscontrol /tmp/sys_log 0
 5546     1 root     S     674m  70%   0% /usr/sbin/upnpd eth0 eth0
 3597     1 root     S    34476   4%   0% /usr/bin/redis-server *:6379
22586 16877 admin    R     3272   0%   0% top
I see that syscontrol is occupying a small amount of memory...

Code: Select all

admin@TNAS-016E80:~$ cat /proc/swaps
Filename                                Type            Size    Used    Priority
/dev/md8                                partition       1998780 0       -1
There is some swap, but I doubt it's on that drive... (which is md1/ sdd, sde )

The pulsing is really quite loud and annoying...

Code: Select all

admin@TNAS-016E80:~$ cat /sys/class/block/md1/stat
   15179        0   266554        0     5376        0  3512936        0        0        0        0
admin@TNAS-016E80:~$ cat /sys/class/block/md1/stat
   15179        0   266554        0     5378        0  3514984        0        0        0        0
admin@TNAS-016E80:~$ cat /sys/class/block/md1/stat
   15179        0   266554        0     5378        0  3514984        0        0        0        0
admin@TNAS-016E80:~$ cat /sys/class/block/md1/stat
   15179        0   266554        0     5380        0  3517032        0        0        0        0

with 'lsof' which is available, I see samba has that drive open...

Code: Select all

root@TNAS-016E80:/home/admin# lsof |grep md1
7223    /usr/sbin/samba4_multicall      /mnt/md1/appdata2
7223    /usr/sbin/samba4_multicall      /mnt/md1/appdata2
7223    /usr/sbin/samba4_multicall      /mnt/md1/appdata2
but I do have one system with the shares connected, but it's not sending anything to the drive...

Re: Drives pulsing every second or so like a heartbeat

Posted: 17 Aug 2023, 09:55
by TMroy
Your drives should be doing array sync, you shouldn't bother with it.

Re: Drives pulsing every second or so like a heartbeat

Posted: 19 Aug 2023, 09:29
by d3x0r
Okay, it stopped by the next afternoon... but when I had first installed the drives it was ticking away for more than a week, until I wrote some data to it and it immediately stopped.

Thanks, the problem did go away on its own.