Jump to content

steffenmand

Members
  • Posts

    418
  • Joined

  • Last visited

  • Days Won

    20

Posts posted by steffenmand

  1. Recently got 10 gbit internet with my server, however are seeing that the maximum size of 20 MB in blocks are causing speeds to be limited around 40 MB/s due to the response time delays pr. thread

    Would love to request that it gets possible to increase the block size further, to acommodate increasing internet speeds over the coming years.

    maybe a 50 MB block size and ofc. a warning showcasing that high internet speeds are required (just like with the minimal download size)

  2. 10 hours ago, Christopher (Drashna) said:

    That depends on the cause, specifically.  Still not sure why you're seeing such high usage though...

    It *may* be possible to throttle the usage, but I'm not sure.

    Also, is this for a single core, or across many/all of the cores?

    Let me flag this, and see if there is something we can do. 
    https://stablebit.com/Admin/IssueAnalysis/27986

     

    Also.... it may be help to "record" this with the Windows Performance Recorder.
    https://docs.microsoft.com/en-us/windows-hardware/test/wpt/windows-performance-recorder

     

    Its all cores! Goes completely bananas and makes the server unresponsive and leaves me fighting to cancel pinning.

    Got 12 drives divided on 5 accounts. 

    Some of them made with pre-release - dont know if that effects it ?

    Will try to do as asked.

    Must add that i didnt have the issue 5 months ago! Something which came with newer versions 

     

     

  3. 9 hours ago, Christopher (Drashna) said:

    THis is probably one of the weird artifacts of how file APIs are handled.  Opening files is CPU intensive.  It usually a tiny spike, but if a lot of files are touched, this can add up. 

     

     

    So the massive CPU usage can not be avoided ? :(

     

    My server ends up a cluster fuck as 12 drives pretty much are pinning all the time as when one finishes a new one begins.

    I got this error on multiple cpu's both xeon , i7-7700 and now i7-8700K :(

  4. Disabled pinning and the problem is gone - so the cpu usage is most definately related to the pinning feature.

    so until solved i will have to live without pinning enabled

  5. 3 hours ago, srcrist said:

    For what it's worth, I see less than 5% usage, generally, on my E3-1245v2. So that might be a bug of some form. I don't think that CPU usage is typical. Certainly not for an 8700k. 

    Do you have multiple drives on multiple accounts ? (i got 12 drives on 5 accounts)

    Usually goes crazy bananas right after it starts pinning.

    The rest of the time it is like you say - less than 5%. Upload, Download whatever - only with the pinning it goes crazy!

  6. Awesome - however i still must note that pinning uses massive amounts of CPU (i7-8700K = 99% in intervals) - does it really require that much computing?

     

    i would still also request that pinning could be limited to X amount of threads. That way pinning doesnt start to use all threads and maybe the CPU usage goes down (although i know that it is not the usage of all threads causing the CPU usage)

  7. Hi,

     

    Was wondering if anyone has issues connecting to the stablebit server? i can't connect no matter what and i have no firewall issues! Happened after upgrading to the newest version (Which can't give a success as the service won't start)

    Launching CloudDrive.Service.exe i see:

     

    "CloudDrive.Service.exe Error: 0 : [Main] Unable to start the StableBit CloudDrive service. Error reading JObject from JsonReader. Path '', line 1, position 23430.
    Error starting service. Error reading JObject from JsonReader. Path '', line 1, position 23430."

    Im thinking that the Service is failing causing the UI not to be able to connect to the service?

     

    UPDATE: Fixed the issue by deleting the data in %PROGRAMDATA%. Apparently it has a hard time overwriting files there - the ProviderSettings.json had all 0's and settings were not there

  8. On 10/12/2018 at 10:12 PM, Christopher (Drashna) said:

    Glad to hear it!  And yeah, we don't get anywhere near the number of bug reports, that we used to. (which is a mixed blessing :D)

    Though, there are other products and services to release now, too. 

    Of course - hope you will look into the pinning issue though - that is a pain in the ass as my computer gets stuck when several drives starts pinning at the same time

  9. Great - thanks :)

     

    This product has become really nice over time and works really well :) Recommending it to everybody i can!

     

    Most of the stuff to be done now is convenience and added providers! Way to go!

  10. I have a couple of requests i have had for a while!

     

    1. Can we mark all attached drives in the main overview with "green" so we can see what has been attached (just like non-attached are yellow). Would be nice to see what is attached - especially when you use multiple accounts and have a hard time remembering which drive is located where (especially if you need to reauthorize)

    2. Queued pinning - make sure only 1 drive can pin at a time - sometimes several drives start pinning and the CPU usage goes to 100% as it goes max thread count on all drives! Would be nice that only 1 drive can pin at a time so i dont have to struggle with cancelling some of the pinnings to control my server.

    3. Limit the amount of threads for pinning! Using 20 threads can quickly suck up all the juice for pinning. Would prefer pinning only using maybe 25% of the threads or maybe 50% so there is still threads free for usage

    4. Some safeguard tool so we can recover our data if Covecube at some point should shut down - currently we are relying on your servers to function!

    Hope some of these things can be incorporated into the product as the product since the beginning has gone to be super awesome - these are the last few challenges i have!

     

     

  11. Still got the issue...

    But it is 100% the pinning causing it. If i manually cancel the pinning then it quickly goes down again.

     

    My guess is that multiple drives start pinning at the same time and they all will run max amount of threads (ie. 15 on mine = 10 drives pinning with 15 threads each = 150 threads)

     

  12. This issue is often causing my drives to unmount due to them not being able to download anything while freezing.

     

    It goes completely crazy at 100% CPU when uploading (and it starts pinning when you upload)

    Computer is unuseable for 1-15 min. depending on the upload size.

     

    Something seriously broke in recent updates - should be seen as a critical issue! What can i do to get you to look into fixing this ?

  13. Never had this pinning issue before i updated after the Google API had some changes (which caused reauthorisation all the time).

    It is related to recent updates! Pinning should ofc. happen, but maybe it should be queued so all drives dont start pinning exactly at the same time ?

  14. 56 minutes ago, steffenmand said:

    After getting Google updated so we dont have to reauthorize again i get massive CPU usage. It can get to so high a load that i cant control my computer for a while when it uploads.

    Didnt have this problem before - did something change ??

    Might be related to pinning ? I see all my drives pinning at the same time when it happens

  15. I have two drives atm.

     

    Provider: Google Drive

    Encryption: No

    Drive Creation: I can't remember but a while ago

    Storage Chunk size: 10 MB

    everything else is the same default settings

     

    Provider: Google Drive

    Encryption: Yes

    Drive Creation: I can't remember one of the newer releases .800~

    Storage Chunk size: 20 MB

    everything else is the same default settings

     

    One note is the encryption drive is currently in a state of chaos where it has errors on it, but i cannot fix the errors partially because half way through the drive dismounts/disconnects and after that the scan fails and i cannot fix it. But i have seen that that original drive i made a while ago has incredible speeds. Close to 300 Mb/s, the encrypted drive gets around 80 Mb/s, I thought maybe it was the encryption issue so i tried to make a new drive no encryption same chunk size. And it only gets about 80 Mb/s as well. I am curious what is causing it to have such good speeds? Is it a case of the errors on the drive so it's overpulling? Just curious. And thought it was an interesting topic of conversation

     

    I also just joined that twitter contest so hope to have a free copy of the cloudDrive! Great product as always

     

    Could it possibly be you who forgot that you set a higher minimal download size ? :) It greatly increases the speeds (i got 400-800 mbit on mine with minimal download of 20 MB)

×
×
  • Create New...