#1
|
|||
|
|||
JDownloader runs like ass
so JD has been running like absolute ass for several weeks now. it is so slow and laggy. everything I do it has these delays and momentary freezes. scrolling, switching tabs, right clicking, every action has a delay.
I'm getting very annoyed because it also seems to be affecting download speeds I have been using JD2 since late 2014 and my linklist has nearly 1.3 million links in it. my linklist (downloadList998393.zip) file is nearly at its millionth iteration and is almost 800MB in size so I am sure this plays a role but it has almost suddenly gotten really bad the last month. I have found that killing JD with the taskmanager while its running can sometimes make it much more stable the next time I start it but it doesn't always work. JD2 is typically having high CPU usage and uses 4.5GB of RAM Is there something wrong with my instance of JD2, if not any plans to make it more stable? I think JD2 needs an overhaul of how it saves the linklist. Most of my linklist never changes and is there for archival and history purposes. JD doesn't need to read and load most of this data or constantly resave it. Maybe JD needs an archive tab where old links can be saved to its own separate file and only loaded when needed. it would be nice to have different labeled archives 24.06.21 11.28.56 <--> 24.06.21 11.43.27 jdlog://5693825302851/ |
#2
|
||||
|
||||
@fuknuckle: JDownloader internas are not designed for JDownloader being used as an archive/history.
That's the simple answer to this. There is no dedicated effort to make it lightweight use/handling for archived/history links. For JDownloader it's just a normal link that you might enable/download at any time. And it highly depends on the link/site if a link requires more or less memory for internal representation. You can set higher delays to write modified list to disk, see Settings-Advanced Settings https://board.jdownloader.org/showpo...09&postcount=4 eg, set DownloadController.minimumsavedelay to 300000 DownloadController.maximumsavedelay to 900000 So JDownloader will wait minimum of 5 minutes after last modifiation to write to disk, but max 15 mins. It will write the changes to disk independent from those settings. Your JDownloader instance is limited to 3.6gb of heap memory MaxMemory=3806855168bytes (3630Megabytes) You can always check memory usage/allocation in about dialog of JDownloader, see https://support.jdownloader.org/Know...tion-directory When the first(used) and second(allocated) number are *near* the third(max), then it's time to remove links from list or incease available max heap. You should modify the .vmoptions file and provide more memory, eg, remove any -Xmx and -Xms line and add -Xmx6g
__________________
JD-Dev & Server-Admin Last edited by Jiaz; 25.06.2021 at 18:52. |
#3
|
||||
|
||||
Quote:
But it's not that simple because most ppl want this *archive/history* to work like the current system, so JDownloader informs/shows that you've already a file or not. So this implementation needs to be *fast* and consume less memory than current *workaround* with keeping links in list. Of course you could just dump those into a database, but this would just create another dependency that might not be available in all supported environments and ppl easily forget that databases are only *fast* if the index mostly fits into memory.
__________________
JD-Dev & Server-Admin Last edited by Jiaz; 25.06.2021 at 19:05. |
#4
|
||||
|
||||
I also have ideas to reduce write pressure and file size when writing lists to disk but that work is not yet done. Just some basic proof of concept.
__________________
JD-Dev & Server-Admin Last edited by Jiaz; 25.06.2021 at 19:06. |
#5
|
||||
|
||||
It would help me to get hands on such a large/long list for working/testing on real life data. Maybe you can upload it to mega and send me the link to support@jdownloader.org ?
__________________
JD-Dev & Server-Admin |
#6
|
|||
|
|||
hey so its been a while since I replied. I solved my issue by doing what you suggested but I was confused for a long time because that vmoptions file didn't exist but after some more searching I got it figured out
It would be really nice if in the future you can increase JD's heap memory in advanced settings you also mentioned that archiving links is a feature that others have requested and that you are working on a proof-of-concept for reducing linklist (and hopefully memory) size. I would like to help anyway I can. I am not a developer but I would be willing to test these features and considering I am a heavy/power user, that might be beneficial to you as for you last request. I have not forgotten but I do have reservations. my linklist contains years worth of download history from many files hosters. Is there any data in the linklist that can identify me by the accounts used to download or any other information that could hurt my privacy? |
#7
|
||||
|
||||
@fuknuckle: didn't find time yet, will read/answer by next week
__________________
JD-Dev & Server-Admin |
|
|