Hello, I am an avid user of Everything and currently using version 1.4.1.1024. As the number of files increases, I have noticed a significant slowdown in the update speed following file modifications. This issue becomes especially noticeable when dealing with around 30 million files. Although I haven't extensively tested version 1.5a, I am curious if there are any improvements in the update speed in response to file modifications in this version. When a large number of files are modified, Everything's window tends to freeze, and I have to wait a considerable amount of time before searches are possible again.
Additionally, when the number of files grows, the file monitor seems to limit the number of files it can update per second. Currently, I have around 100 million files, and handling this with a single instance of Everything has become impractical due to the slowdowns. To mitigate this, I have been using three separate instances of Everything, which has somewhat alleviated the bottleneck, but this feels more like a temporary workaround than a permanent solution.
I actively use the Everything SDK, but I find it lacking the functionality to specify instances. I've modified the SDK code myself to create such a function, but an official update would be highly appreciated.
Moreover, integrating Everything with NAS devices like Synology is cumbersome since the only option seems to be adding network paths and relying on Daily Scans for indexing. If Everything Server, introduced in 1.5a, could be enabled, it would allow for a more efficient use of Everything without the overhead of Daily Scans.
Questions:
Are there any improvements in the update speed in response to file modifications in Everything version 1.5a compared to 1.4.1.1024?
Is there any way to overcome the file monitor update limit per second when dealing with large numbers of files?
Are there plans to officially add the functionality to specify instances in the Everything SDK?
How can Everything Server be enabled on NAS devices to avoid the overhead of Daily Scans?
Improving Everything Search Performance for Large File Sets
Re: Improving Everything Search Performance for Large File Sets
Everything 1.5 is noticeably faster than 1.4. Furthermore: most updating is now done in the background, which means that searching during that time is still possible.
Everything loads its entire database in RAM memory. 1 million items (files/folders) take about 100 MB RAM, so 30 million files will consume around 3 GB memory. Make sure your system has enough free memory available as swapping to disk will make Everything extremely slow (and use the 64-bit version of Everything as the 32-bit version cannot address that amount of memory)
The SDK for Everything 1.5 will support instances when released (not during the alpha stage)
Everything Server is a Windows application; it will not run on most NAS devices.
The rescan schedule can be changed from daily to every x hours or even every x minutes.
Everything loads its entire database in RAM memory. 1 million items (files/folders) take about 100 MB RAM, so 30 million files will consume around 3 GB memory. Make sure your system has enough free memory available as swapping to disk will make Everything extremely slow (and use the 64-bit version of Everything as the 32-bit version cannot address that amount of memory)
The SDK for Everything 1.5 will support instances when released (not during the alpha stage)
Everything Server is a Windows application; it will not run on most NAS devices.
The rescan schedule can be changed from daily to every x hours or even every x minutes.