Anybody tried an external SSD?

[graemev]

I wonder if there is any mileage in plugging in an SSD into a USB port, then using links or otherwise direct all the logs etc onto the SSD?
(actually mounting over /var/log etc might be a better starting point?)

[dolbyman]

Remember QTS is an embedded system, many of those fiddle settings would be changed upon reboot or firmware update. So messing around with this could end in tears (even if not immediately)

you probably also want to read this

viewtopic.php?t=130788

[graemev]

Indeed it was reading that post that prompted my comment.

I could see he was going to a lot of effort to free up a bay to use, I though why not an extra external disk?

You'd need to ensure e.g /var/log was on the SSD before any process opened it's log , otherwise it'd just continue to use the inode it had already opened.

...I've not thought this through yet , but simply:

Code: Select all

mount /dev/sse1  /var/log

Early on.

I see /etc/fstab is rather bogus looking , so I assume QTS does it's own mounts, then a pivot_root (2)
so where to do it could be a problem but (assuming there's a safe place to do so) simply mounting other filesystems in the tree is not very dangerous ... like /boot and /home on regular systems. (you can do the mount or not, system will still work ...you might be missing some files unless you take steps)


(just about the shut my NAS down for the evening as it chattering (/var/log/networking/* ) is getting
ridiculous.

[Pete842]

In my QNAP NAS (TS-h973AX) I already have 6 SSDs NVMe and SATA mounted internally that would be big enough to run the OS and even VMs from without periodically accessing the HDDs. But even on this officially SSD equipped hardware, HDDs do not spin down because the OS periodically writes to the HDDs.
The way QuTS-Heroe works makes it extremely difficult to relocate the system or parts of it to the SSD. And any attempts likely get reverted with each update of the OS.
And yes, shutting down the NAS over night is not a viable workaround for me too. Especially as my TS-h973AX takes about 10 minutes to boot. My M1 Mac takes more like 20 seconds!

I‘ve setup a poll how people deal with the problem. Be welcome to add your vote.

[graemev]

In my case (YMMV) the writes are all to the logs (/var/log/*) and they look very much like somebody has shipped the development version of the code (with debugging turned on) . I've not working this though, just floating the idea. But if I could just get all those logs on a single (external) SSD I think my disk would spin down. I did have them spinning down a while back. It was really obvious because as soon as it touched the computer in the morning they would all spin up in sequence.

The trivial (not thought though) idea would be to add an entry to init.d to simply mount over /var/log , so i should survive upgrades.

[graemev]

This is getting more and more of an issue. This afternoon , the fan on my TS451+ went into "higher speed" while nothing was accessing the drive, just the internal debugging constant disk writes.
Just noticed another firmware update has arrived hopefully they've address this.
(they seem to waste lots of effort on features nobody can use, in part because of these basic flaws. Time would be better spent getting the basic functionality working)

[dolbyman]

I saw this article a while ago .. a nice compromise of kicking the drives off the system but doing a daily sync to all of them

https://alex3305.github.io/home-assista ... -spindown/

Not that I would go that route, but interesting read

[graemev]

interesting, going to need a bit of thought (and an SSD ...damm only my front USB port is spare ...maybe not, they stopped printer support)

Code: Select all

[graeme@qnap451plus network]$ cat /proc/mdstat                                                                                                        
Personalities : [linear] [raid0] [raid1] [raid10] [raid6] [raid5] [raid4] [multipath] 
md1 : active raid6 sdb3[5] sdc3[4] sdd3[7] sda3[6]
      7794126848 blocks super 1.0 level 6, 512k chunk, algorithm 2 [4/4] [UUUU]
      
md322 : active raid1 sdc5[3](S) sdd5[2](S) sda5[1] sdb5[0]
      7235136 blocks super 1.0 [2/2] [UU]
      bitmap: 0/1 pages [0KB], 65536KB chunk

md256 : active raid1 sdc2[3](S) sdd2[2](S) sda2[1] sdb2[0]
      530112 blocks super 1.0 [2/2] [UU]
      bitmap: 0/1 pages [0KB], 65536KB chunk

md13 : active raid1 sdb4[33] sdc4[32] sdd4[35] sda4[34]
      458880 blocks super 1.0 [32/4] [UUUU____________________________]
      bitmap: 1/1 pages [4KB], 65536KB chunk

md9 : active raid1 sdb1[0] sdc1[3] sdd1[2] sda1[1]
      530048 blocks [4/4] [UUUU]
      bitmap: 0/65 pages [0KB], 4KB chunk

unused devices: <none>
[graeme@qnap451plus network]$ pwd                                                                                                                    
/var/log/network
[graeme@qnap451plus network]$ df -h .
Filesystem                Size      Used Available Use% Mounted on
none                    400.0M    307.2M     92.8M  77% /

OK I guess I need to understand how that "none" relates to one of the above

Code: Select all

[graeme@qnap451plus network]$ mount
none on /new_root type tmpfs (rw,mode=0755,size=409600k)
/proc on /proc type proc (rw)
/proc on /proc type proc (rw)
devpts on /dev/pts type devpts (rw)
sysfs on /sys type sysfs (rw)
tmpfs on /tmp type tmpfs (rw,size=64M)
tmpfs on /dev/shm type tmpfs (rw)
tmpfs on /share type tmpfs (rw,size=16M)
/dev/sde5 on /mnt/boot_config type ext2 (rw)
tmpfs on /mnt/snapshot/export type tmpfs (rw,size=16M)
/dev/md9 on /mnt/HDA_ROOT type ext3 (rw,data=ordered)
cgroup_root on /sys/fs/cgroup type tmpfs (rw)
none on /sys/fs/cgroup/memory type cgroup (rw,memory)
/dev/mapper/vg1-lv1312 on /mnt/pool1 type ext4 (rw,data=ordered,barrier=1,nodelalloc)
/dev/mapper/cachedev1 on /share/CACHEDEV1_DATA type ext4 (rw,usrjquota=aquota.user,jqfmt=vfsv1,user_xattr,data=ordered,data_err=abort,delalloc,nopriv,nodiscard,noacl)
cpu on /sys/fs/cgroup/cpu type cgroup (rw,cpu)
/dev/md13 on /mnt/ext type ext4 (rw,data=ordered,barrier=1,nodelalloc)
tmpfs on /samba_third_party type tmpfs (rw,size=32M)
tmpfs on /share/.quftp_client type tmpfs (rw,size=1m)
tmpfs on /share/CACHEDEV1_DATA/.samba/lock/msg.lock type tmpfs (rw,size=48M)
tmpfs on /mnt/ext/opt/samba/private/msg.sock type tmpfs (rw,size=16M)
nfsd on /proc/fs/nfsd type nfsd (rw)
/tmp on /share/CACHEDEV1_DATA/.qpkg/MediaSignPlayer/CodexPackExt/tmp type none (rw,bind)
hdsfusemnt on /share/CACHEDEV1_DATA/.qpkg/MediaSignPlayer/CodexPackExt/share type fuse.hdsfusemnt (rw,nosuid,nodev,allow_other)
hdsfusemnt on /share/CACHEDEV1_DATA/.qpkg/HD_Station/share type fuse.hdsfusemnt (rw,nosuid,nodev,allow_other)

So I guess this is what he's looking at .... don't see my logs there...

/dev/md9 on /mnt/HDA_ROOT type ext3 (rw,data=ordered)

Code: Select all

[graeme@qnap451plus network]$ ls -l /mnt/HDA_ROOT  
total 124
-rw-r--r-- 1 admin    administrators 45056 2015-11-17 19:55 ds.db
drwx------ 2 admin    administrators 16384 2012-02-17 04:40 lost+found/
-rw-r--r-- 1 admin    administrators 15218 2017-10-16 18:45 md_backup_2017-10-16_18.37.42
-rw-rw-rw- 1 admin    administrators 15271 2019-04-04 12:45 md_backup_2019-04-04_12.45.25
-rw-rw-rw- 1 admin    administrators 15271 2019-04-04 14:42 md_backup_2019-04-04_14.42.53
-rw-rw-rw- 1 admin    administrators  3456 2019-04-17 18:12 md_backup_2019-04-17_18.12.18
-rw-rw-rw- 1 httpdusr administrators     0 2014-02-23 21:08 photo_scan.record
drwxr-xr-x 2 admin    administrators  4096 2023-02-17 16:25 ssl_lib/
lrwxrwxrwx 1 admin    administrators    24 2020-04-11 12:48 twonkymedia -> /mnt/ext/opt/twonkymedia
drwxr-xr-x 2 admin    administrators  4096 2021-04-17 18:30 update/
drwxr-xr-x 3 admin    administrators  4096 2023-02-17 16:43 update_pkg/
[graeme@qnap451plus network]$ ls -ld /hd_root_tmp
lrwxrwxrwx 1 admin administrators 13 2023-02-17 16:33 /hd_root_tmp -> /mnt/HDA_ROOT/
[graeme@qnap451plus network]$ 

Interesting running a find(1) from root causes no more disk activity than these damm debug logs . BTW the latest firmware is even worse ...came here to see how to raise a ticket.

Code: Select all

[graeme@qnap451plus network]$ tail default.log
2023-02-17 20:28:16.909 139712197199616 DEBUG [ethtool_utils.cpp:459] int check_physical_format(int, int, int) >>> Exit 0
2023-02-17 20:28:16.909 139712197199616 DEBUG [NetworkInterface.cpp:2259] int NetworkInterface::check_config_value() >>> Exit 0
2023-02-17 20:28:16.909 139712197199616 DEBUG [NetworkInterface.cpp:630] int NetworkInterface::NI_member_to_struct(NI_member*) >>> Enter
2023-02-17 20:28:16.909 139712197199616 DEBUG [PhysicalInfo.cpp:348] int NIC_Info::get_NIC_info(PHY_INFO&) >>> Enter
2023-02-17 20:28:16.909 139712197199616 DEBUG [PhysicalInfo.cpp:365] int NIC_Info::get_NIC_info(PHY_INFO&) >>> Exit 0
2023-02-17 20:28:16.909 139712197199616 DEBUG [PhysicalInfo.cpp:425] int PCIE_Info::get_PCIE_info(PHY_INFO&) >>> Enter
2023-02-17 20:28:16.909 139712197199616 DEBUG [PhysicalInfo.cpp:430] int PCIE_Info::get_PCIE_info(PHY_INFO&) >>> Exit 0
2023-02-17 20:28:16.909 139712197199616 DEBUG [NetworkInterface.cpp:692] int NetworkInterface::NI_member_to_struct(NI_member*) >>> Exit 0
2023-02-17 20:28:16.909 139712197199616 DEBUG [NetworkInterface.cpp:6373] int check_interface_config_ip6_gw(char*) >>> interface ip6_type: DISABLED
2023-02-17 20:28:16.909 139712197199616 DEBUG [NetworkInterface.cpp:6393] int check_interface_config_ip6_gw(char*) >>> Exit 0

Looks like is it's damm near doing method entry exit tracing ....do these guys do NO QC before shipping the code?

[graemev]

Well this is distinctly odd ... I got an update yesterday (24Apr2024) that @Muriel47 had added post here. I read their post yesterday. Today it is gone and indeed there is no trace of @Muriel47 ???? Not sure what is going on? Anyhow, I have put in many man months of work on this and about 3 weeks ago switched over to my "new system". I was waiting to see how stable it was first (seems fine).

Rather than break my TS451+ (16GB RAM, Intel CPU) I implemented my changes on my TS412 (old ARM chip 256Mb RAM) . It's running (almost) totally silent for most of the day. The fan occasionally spins up on it's lowest setting (Silent) but all the disk are stopped and it draws around 7W. It has the latest "bookworm" running and I'm able to stream 4K video just fine ...so I think there is enough CPU "grunt" here. There are some issues with the very small amount of RAM but I'm not going to address this as I plan to switch to a simple "Orange Pi" with 8GB RAM and some NVMe .

As an added bonus, it's diskless (flashless?) boots from a TFTP server , so I don't need to re-flash the NOTR memory ..I had bad experiences here

(viewtopic.php?t=172343) but @NickJ_UK saved the day with a replacement M.board.

[OneCD]

Well this is distinctly odd ... I got an update yesterday (24Apr2024) that @Muriel47 had added post here. I read their post yesterday. Today it is gone and indeed there is no trace of @Muriel47 ???? Not sure what is going on?

@Muriel47 turned-out to be a spambot, so they have been banned and their posts removed.