[SCRIPT] create-autorun.sh

[RichJacot]

Yes, it has remained through all of my reboots. hmmm... It shouldn't?

[/share/CACHEDEV1_DATA/.system/autorun] # cd /
[/] # ll
total 4.0K
drwxr-xr-x 22 admin administrators 540 2020-05-06 17:27 ./
drwxr-xr-x 22 admin administrators 540 2020-05-06 17:27 ../
-rw-r--r-- 1 admin administrators 135 2020-05-06 17:25 1
drwxr-xr-x 2 admin administrators 2.1K 2020-05-06 17:22 bin/
drwxr-xr-x 15 admin administrators 21K 2020-05-06 17:35 dev/
drwxr-xr-x 34 admin administrators 3.2K 2020-05-06 17:35 etc/
drwxr-xr-x 2 admin administrators 40 2020-05-06 17:22 flashfs_tmp/
lrwxrwxrwx 1 admin administrators 13 2020-05-06 17:22 hd_root_tmp -> /mnt/HDA_ROOT/
drwxr-xr-x 4 admin administrators 80 2020-04-09 15:57 home/
drwxr-xr-x 7 admin administrators 4.4K 2020-05-06 17:23 lib/
lrwxrwxrwx 1 admin administrators 3 2020-05-06 12:21 lib64 -> lib/
lrwxrwxrwx 1 admin administrators 11 2020-05-06 12:21 linuxrc -> bin/busybox*
drwx------ 2 admin administrators 40 2020-04-09 15:58 lost+found/
drwxr-xr-x 13 admin administrators 280 2020-05-06 17:24 mnt/
drwxr-xr-x 2 admin administrators 40 2020-05-06 12:21 new_root/
drwxr-sr-x 2 admin administrators 200 2020-05-06 17:27 opt/
lrwxrwxrwx 1 admin administrators 19 2020-05-06 12:21 php.ini -> /etc/config/php.ini
drwxrwxrwx 2 admin administrators 40 2020-05-06 17:27 plex.transcoding.tmp/
dr-xr-xr-x 340 admin administrators 0 2020-05-06 12:21 proc/
drwxr-xr-x 2 admin administrators 240 2020-05-06 17:25 root/
drwxr-xr-x 2 admin administrators 40 2020-05-06 17:25 rpc/
drwxr-xr-x 2 admin administrators 9.6K 2020-05-06 17:26 sbin/
drwxrwxrwt 31 admin administrators 800 2020-05-06 17:25 share/
dr-xr-xr-x 12 admin administrators 0 2020-05-06 11:21 sys/
drwxrwxrwx 18 admin administrators 2.5K 2020-05-06 18:04 tmp/
drwxr-xr-x 9 admin administrators 200 2020-04-09 14:50 usr/
drwxr-xr-x 16 admin administrators 740 2020-05-06 18:04 var/
[/] #

[OneCD]

Yes, it has remained through all of my reboots. hmmm... It shouldn't?

Only that yours would be the first. QTS is an embedded OS. It's not like a regular Linux desktop or server distro. It has a preset and read-only filesystem image which is loaded on boot and expanded into RAM. So, user-created directories from the root filesystem are not saved or restored.

How are you checking your new tmpfs has been created?

What's the output from:

Code: Select all

df

... ?

[RichJacot]

Yes, df. Let me remove my script and reboot to see if it stays. I haven't looked without my script in place.

I'll send you the outputs shortly.

Sent from my SM-N975U using Tapatalk

[OneCD]

I wonder if it's being created by something else after autorun.sh is executed?

Try your script with these mods:

Code: Select all

#!/bin/bash
TARGET_PATH=/plex.transcoding.tmp
[[ ! -d $TARGET_PATH ]] && { mkdir -p $TARGET_PATH; echo "created target" ;} || echo "target already exists"
/bin/mount -t tmpfs -o size=16G tmpfs $TARGET_PATH; echo "a:$?"
/bin/touch /share/CACHEDEV1_DATA/.system/autorun/scripts/plxran; echo "b:$?"

[RichJacot]

Sorry, to many users now. I'm going to have to do this in the morning.

Sent from my SM-N975U using Tapatalk

[RichJacot]

Okay, I removed my script and sure enough my /plex.transcoding.tmp wasn't there. Interesting but I hadn't rebooted without my script in place and every time I looked my /plex.transcoding.tmp dir was there.

I made the changes you suggested and now it appears to be working perfectly now. /plex.transcoding.tmp is there, I can see it with df and below is the autorun.log

[/share/CACHEDEV1_DATA/.system/autorun/scripts] # cat /var/log/autorun.log
Thu May 7 06:27:14 CDT 2020 ----- running autorun.sh -----
Thu May 7 06:27:14 CDT 2020 - /share/CACHEDEV1_DATA/.system/autorun/scripts/PlexTranscodeRAMDrive.sh
created target
a:0
b:0
[/share/CACHEDEV1_DATA/.system/autorun/scripts] #

I ran a couple of streams earlier this morning that needed to be transcoded and I could watch /plex.transcoding.tmp grow.

Thank you Sir!

[OneCD]

I ran a couple of streams earlier this morning that needed to be transcoded and I could watch /plex.transcoding.tmp grow.

Good work.

So, your script can now be reduced to:

Code: Select all

#!/bin/bash
TARGET_PATH=/plex.transcoding.tmp
[[ ! -d $TARGET_PATH ]] && mkdir -p $TARGET_PATH
/bin/mount -t tmpfs -o size=16G tmpfs $TARGET_PATH

[RichJacot]

Roger that and thanks again!

[MikeLagit]

I'm using Entware with mlocate installed for indexing files for fast finds. I need to prune out the snapshots directory as it creates a ton of duplicates I don't care about. If I edit the /etc/updatedb.conf, it is overwritten at reboot. I don't see a way to implement a permanent version of the config file (correct me if I am wrong), so I resorted to this supercool tool from OneCD. I have it setup, but it's still getting overwritten at boot. Is this an ordering problem? I tested the script below and it adds the prunepath when run it manually, I just cant get it to stick after a reboot. Please see below and let me know if you have suggestions.

[~] # cat /opt/etc/updatedb.conf
PRUNENAMES="/.AppleDB/.AppleDouble/.AppleDesktop/:2eDS_Store/Network Trash Folder/Temporary Items/TheVolumeSettingsFolder/.@__thumb/.@__desc/:2e*/.@__qini/.Qsync/.@upload_cache/.qsync/.qsync_sn/.@qsys/.streams/.digest/.*/"

The above output doesn't have my additional line and should look like this:

[~] # cat /opt/etc/updatedb.conf
PRUNENAMES="/.AppleDB/.AppleDouble/.AppleDesktop/:2eDS_Store/Network Trash Folder/Temporary Items/TheVolumeSettingsFolder/.@__thumb/.@__desc/:2e*/.@__qini/.Qsync/.@upload_cache/.qsync/.qsync_sn/.@qsys/.streams/.digest/.*/"
PRUNEPATHS="/mnt/snapshot"

If I look at my logs it says the script was run:

[~] # cat /var/log/autorun.log
Thu May 14 12:15:02 CDT 2020 ----- running autorun.sh -----
Thu May 14 12:15:02 CDT 2020 - /share/CACHEDEV1_DATA/.system/autorun/scripts/010-mlocate.sh

Running it manually does add the prunepaths to the /opt/etc/updatedb.conf (tested it manually to make sure script is valid and executable).

[~] # cat /share/CACHEDEV1_DATA/.system/autorun/scripts/010-mlocate.sh
#!/bin/bash
echo "PRUNEPATHS=\"/mnt/snapshot\"" >> /opt/etc/updatedb.conf

Please let me know what I could be doing wrong. Is this a case for the RunLast qpkg?

[OneCD]

Is this a case for the RunLast qpkg?

RunLast would be one way to do it. But the problem is that Entware can take some time to complete startup reintegration while the NAS is booting. RunLast may execute your script while Entware is unavailable. Which means you would need to introduce a wait-and-check poll to test for Entware availability over about a 10 minute period. This works but is not ideal.

A neater solution may be to use the init.d system included with Entware. This is just like the ancient init.d system included with QTS, but works with Entware installed IPKGs.

So, your script would be placed into [/opt/etc/init.d] and renamed to start with an 'S'. It will then run when Entware's startup reintegration is complete.

[MikeLagit]

I don't see any other S scripts in that dir. Is that the correct location? You're saying I can just move that script into that location and it should execute it?

[/opt/etc/init.d] # ll
total 16K
drwxr-xr-x 2 admin administrators 4.0K 2018-06-07 19:09 ./
drwxr-xr-x 6 admin administrators 4.0K 2020-04-28 17:03 ../
-rw-r--r-- 1 admin administrators 2.8K 2018-03-12 14:22 rc.func
-rwxr-xr-x 1 admin administrators 966 2018-03-12 14:22 rc.unslung*
[/opt/etc/init.d] #

[OneCD]

I don't see any other S scripts in that dir. Is that the correct location?

Which only means there are no currently installed IPKGs with init scripts. The only ones requiring init scripts are daemons.

You're saying I can just move that script into that location and it should execute it?

That's the theory.

Try it and see. Add your script to that location, and rename it to prefix an 'S'. Then 'stop' - wait - then 'start' the Entware QPKG in your App Center (no need to reboot the NAS). This will run the same package reintegration process as when QTS is booting. When this completes, your script should be executed.

[MikeLagit]

Worked like a charm! Thanks you for the help!!

[Qmann]

I need to figure out the best way to sticky permissions and group on /dev/dri. Would autostart.sh be the best place to do it? Here is the issue:

For Plex hardware transcoding, I need the /dev/dri to be in the video group, and readable by the plex user which is in docker. I have the /dev/dri passed to the docker and no problem getting it to work, but after each reboot /dev/dri is set back to admin:administrators. After each reboot, I need to issue these commands to get hw transcode going again:

chmod ug+rwX -R /dev/dri
chgrp -R video /dev/dri
docker restart linuxserver/plex (so it picks up the changes)

What is the best fix here? autostart.sh would get things changed on the system, but somehow it would have to come before docker get started to avoid another restart.

SMP:

* New order *
(01) (A) Entware
(02) (A) NvKernelDriver
(03) (A) netmgr
(04) (A) NodeJS
(05) (A) QJDK8
(06) (A) container-station
(07) (A) Qapache
(08) (A) MySQL8
(09) (A) MultimediaConsole
(10) (A) HybridBackup
(11) (Φ) helpdesk
(12) (Φ) Qboost
(13) (Φ) ResourceMonitor
(14) (Φ) QKVM
(15) (Φ) PlexPy
(16) (Φ) QSubSonic
(17) (Φ) NotificationCenter
(18) (Φ) SSDLaboratory
(19) (Φ) SecurityCounselor
(20) (Φ) browser-station
(21) (Φ) iPerf3
(22) (Φ) LicenseCenter
(23) (Φ) QcloudSSLCertificate
(24) (Ω) MalwareRemover
(25) (Ω) SortMyQPKGs

Thanks,
Qmann

[OneCD]

What is the best fix here? autostart.sh would get things changed on the system, but somehow it would have to come before docker get started to avoid another restart.

This may work, if 'autorun.sh' is executed before Container Station. Try it and see.