I tired the suggestion in the other thread and restarted odrive but my machine is just not usable when odrvie is running anymore, not sure what happened, it used to be fine left running 24/7. The only recent change I made was installing the latest Paragon drivers because the built-in ntfs is not reliable nor supported.
However, I can say that when I quit odrive, my machine runs nice and snappy again even with Paragon running. Not sure if it helps to know but it seems that when odrive is consuming a lot of CPU, it appears to do it in pair with windowServer. When I shut odrive down, windowServer quiets down and the machine is usable.
Yes an NTFS volume, single synced connection to Amazon Cloud Drive with approx. 140Gb used on a 3Tb external drive connected via USB3.0. No other cloud accounts in use with odrive.
The odrive folder is on the root of the drive in a folder named “odrive”. Everything should be in sync as I haven’t changed files in days but I still see constant usage in Activity Monitor. I can also see activity in Finder within the odrive folders if I sit and watch for a while, previously shown “in sync” blue icons turn pink again then back to blue as it cycles through each folder. It doesn’t ever appear to rest.
Just reviewed the docs and it does not appear to accurately reflect the client I have.
First, it mentions and shows in the first image a “blue” badge background around the client’s icon in the tray supposedly signifying completion of sync. I have never seen this. The icon is either light grey or pink and the background is always transparent.
Second, the options shown on the screenshots are clearly of a MAC client but are not the same as my MAC client. Things like:
Manage Links
Syncing (6)
Stop Sync
Throttling
Third, I get the blue/pink icons within my odrive folders but only sometimes get the (right-click) context menu with odrive options. Sometimes I do, sometimes I don’t. For example, right now I have blue/pink icons but no right-click options to sync/unsync etc.
Lastly, when I choose to disable auto-sync, I still see sync activity in the Finder, showing pink sync icons as it continues to go through folders. My CPU is still spiked and Activity Monitor still shows odrive process using 80-120% even 10 minutes after disabling auto-sync. The only way to free my CPU is to exit the client unfortunately.
I basically have to use odrive only at night when I am sleeping and keep it off while I work during the day.
Hi @p_ingram3541,
I’ve moved this into its own thread for better tracking.
It looks like the screenshots from some items need to be updated, although the functionality is the same.
The periodic pink on the odrive menu icon indicates periodic scanning that odrive needs to do to make sure things are all in sync between the odrive folders and the remote storage. When auto-sync is disabled, odrive will not upload or download any changes automatically, but the background scanning will still take place so that odrive understands what the changes are.
We don’t officially support moving the odrive folder to an external volume. The reason for this is that it can lower performance and cause additional overhead, which may account for some of the behavior you are seeing.
There is always going to be some overhead when odrive performs its periodic scans. The impact of this overhead is dependent on a few things:
The number of objects being tracked. The number of folders, especially
The location of the odrive folder. External volumes will have poorer performance and cannot take advantage of some optimizations.
The type of storage and number of storage links you have connected.
If you send a diagnostic from the odrive menu, I can take a look to see if anything else sticks out.
Thanks, I’ll submit the diagnostics but I must also mention that I’ve been using odrive for several years now on this machine with the same external drive. Everything you mention would be a great consideration if I didn’t already have good, long-term performance with odrive using this method.
Nothing significant in the storage itself has changed recently related to size/qty files, less than 1% change in file ratio of approx. 140gb. Lot’s of web stuff so yes very folder intense but has always been that way and again, no more than a 1% variance per day of use.
The performance difference has not been gradual as if to suggest a growing file system as the issue, it was more immediate going from one day, no known performance problems to another of Hey, I can’t use my machine, what is going on type problem. I am very careful of what I download and use and watch everything carefully and was pretty shocked to find odrive as source of my slowdown. Now just trying to find out what happened to odrive or what is happening as a result of something else…like Paragon.
When I shut down Paragon (not sure if something is still changed/running behind the scenes) the drives remain mounted and browsable and odrive still runs wild even when stopped and restarted (BTW, odrive finder extension never quits without force). And again, right click menu is still missing the odrive shortcuts but I do see the blue/pink folder icons in Finder folder structure.
Hi @p_ingram3541,
I took a look at the diagnostic, but it looks like it was sent shortly after starting odrive, so it doesn’t really show a whole lot. Initial scan on startup will always spike CPU as it runs through all of the folders to make sure nothing has changed while its been off.
As far as changes, we haven’t change the sync behavior in quite some time, so nothing has changed on the odrive side of things. I believe Paragon installs a low-level driver, too, so it is probably still running even if the Paragon software is closed.
Are these CPU spikes you are seeing periodic, or constant?
You can expect spikes on startup and then periodically (approximately every 1-1.5 hours) as odrive does a full scan to make sure things are all synced up. During that time the CPU can spike for a few minutes (the length of time depends on the size of your expose odrive folder) as odrive plows through its scan. Collapsing folders with unsync can reduce the scope which will reduce the overall overhead, if that is an option.
Since you said that things were working fine until after installing Paragon, and nothing else has changed, it would seem that Paragon is a likely culprit. You could try uninstalling it completely to see if there is any difference, just to pinpoint that particular possibility.
Ok, that seemed to do the trick. It still spikes the CPU 50-80% but not constantly like the last few weeks. It mostly settles down to less than 5% now which does not seem to impact performance. I think I am back in shape.
This is a big win but also kind of a bummer because I really need reliable NTFS drivers, especially when connecting new external drives and needing read/write access. The /etc/fstab method doesn’t always work and is a pain when a friend hands me a drive that’s never been setup on my system before.
Purchasing Paragon fixed that problem. I’ll open a ticket with them.
Regardless you’ve been great through this whole process and I appreciate how involved you get in helping us even though often we are the ones messing everything up
I appreciate all of the details and the willingness to jump through hoops to figure this out. It sucks that this behavior is happening. I am wondering if it has to do with the filesystem stat calls we are making and it is somehow creating additional overhead when running through the Paragon driver.
We aren’t doing anything special, so it is strange that CPU would spike when performing standard filesystem lookups and operations through Paragon. Please let me know what you end up hearing from Paragon on this.
John did provide some links detailing how I can send him a sysdiagnose file and logs after reproducing the behavior but I’ve already uninstalled Paragon and enjoying my machine being zippy again…
Again individually both software seem to work great it’s just when you get them on the same playground…or maybe just my playground.
I’ll likely get some time to play with this over the weekend.
