Hi,
I'm using SAB from SynoCommunity on a Synology DS212 with CPU@1,6ghz with a time-based usenet provider. For better perfomance and resource managemnt it would be great if the following features could be added.
1) Number of connections per server
Now the CPU runs on 60-98% during downloading with 12 connections to receive full speed (disabled PostProcessing). So if for example I can set the connection to 6 during PostProcessing it could run simultaneously or avoid WebUI cherrypy problems during 12:00-16:00. Other example if you have a slow connection you can set it to a lower # during daytime to save bandwith and resources.
2) Run jobs once
For example schedule a restart of SAB
3) Cleanup history (completed jobs from history etc..)
So the WebUI doesn't need to retrieve the information by accessing it during download time.
4) Clear article buffer
Now when enabled the buffer it doesn't dump its content if all servers are disabled for a while.
5) Change time interval for Watchfolder and max. number of files or minutes
For example I want to scan it between 12:00-16:00 and after that set it to 0 to avoid HDD standby prevention. Or the option to set a scan event for 23:45 read a maximum of 15 files or minutes.
Thanks for your time!
Schedule feature: #connections per server, run restart once
Schedule feature: #connections per server, run restart once
Last edited by waldema on August 19th, 2012, 8:59 am, edited 3 times in total.
Re: Schedule feature: #connections per server, run once
Not all suggestions makes sense to me.
Answers to your suggestions in italic.
I still find most of them just too exotic
1. Why vary the # of connections?
I think it's better to stop downloading during post-processing (already an option)
or define one server twice as primary each with 6 connections
2. What kind of jobs (SABnzbd scheduler is not a crontab replacement)
One shot might be useful, I'll add that to planned changes for 0.8.0
3. Could be useful, but why is this better than a periodic manual cleanup?
It's not like the history is very expensive to keep
The WebUI only asks the data for the current page (unless you're using Classic: don't do that).
4. If you pause, the article cache is dumped.
Some Python implementations on some platform are very reluctant to return memory
to the OS, but this won't fix it.
Schedule a pause right after the last server is disabled
5. You can already choose a scheduled read of the watched folder
Number of files?
Sorry, that's just too exotic.
Maybe you can add some more motivation?
Answers to your suggestions in italic.
I still find most of them just too exotic
1. Why vary the # of connections?
I think it's better to stop downloading during post-processing (already an option)
or define one server twice as primary each with 6 connections
2. What kind of jobs (SABnzbd scheduler is not a crontab replacement)
One shot might be useful, I'll add that to planned changes for 0.8.0
3. Could be useful, but why is this better than a periodic manual cleanup?
It's not like the history is very expensive to keep
The WebUI only asks the data for the current page (unless you're using Classic: don't do that).
4. If you pause, the article cache is dumped.
Some Python implementations on some platform are very reluctant to return memory
to the OS, but this won't fix it.
Schedule a pause right after the last server is disabled
5. You can already choose a scheduled read of the watched folder
Number of files?
Sorry, that's just too exotic.
Maybe you can add some more motivation?