Hi wschmid,
I realize this is a pretty old post but I came across it today while diagnosing a similar issue on our own Vault.
I figured out something really interesting that might give you and other users some insight into what is going on.
TLDR: Files in Vault that have been modified since the last ADMS indexing job may take significantly longer to GET than other files. The slowness is not specifically due to the number or size of the files you are GETting. It's all about the date it was modified.
As you mentioned, we have some folders with LOTS of files, totaling hundreds of megabytes, that we can Get very quickly -- no issues. However, other folders with fewer files and less files size may appear to lock up the Vault client and it takes many minutes to nearly an hour to Get the folder.
It was really puzzling and didn't make a lot of sense...until today I noticed this:
Files that have been modified today will often take much longer to Get.
As long as the files were modified prior to today, we can Get hundreds of them at a time with no performance issues. However, trying to Get even a few of the files that were recently modified results in a crazy delay.
Vault will eventually download them and the client doesn't every actually crash but you will get the 'Not Responding' message and it will sometimes freeze while it is doing the GET.
Why would this possibly be?
Why does Vault care if the file was modified today or not? I'm not sure of that yet but my hunch is that it is due to Indexing. Vault uses SQL Server, as we know. However, it also creates its own local index of that database (referred to as the Lucene index, I believe) and it is MUCH faster for Vault to look in its own index than to have to search in the SQL database index. Currently, we run the Search Index Optimization ADMS task only once per day, offloaded to non-production hours so that it doesn't impact work. In my testing, it is clear that as long the Optimization job has run after the files in question have been modified, those files can be downloaded very quickly with the Get. If they have been modified today, i.e. before the Search Index Optimization job has completed, the Get will complete but it takes significantly longer to do so.
We have users who often need to Get files that have been modified recently by other users, so I'm considering adjusting the interval that I run the Optimization so that it re-indexes these files more quickly. This should speed up the Get considerably.
If anyone else has experienced this behavior and has any other thoughts, I'd love to hear them.
I hope this observation helps someone else who has encountered this puzzling 'Slow Get' issue.
Cheers!