You are viewing limited content. For full access, please sign in.

Discussion

Discussion

Thread blocking for SDK App

SDK
Updated October 21, 2019
posted on October 18, 2019

I have this scheduled task that uses the SDK and loops about 6,600 entries.  It will return maybe 100 or so records and then block the thread consistently for 2 min 30 sec before it will give me another 100.  The whole process takes more than an hour to run.  The line of code where it's blocking is indicated in the attachment.

Anyone know if there's a setting, or something I can do in my code to not make it block?  

       

10-18-2019 6-48-08 PM.png
10-18-2019 6-48-08 PM.png (73.54 KB)
Download
0 0
replied on October 19, 2019

It looks like you are running an expensive search by finding all entries with a certain template and subtracting the "Retention" folder. Does it speed up the process if you exclude the LF:Lookin phrase and have your code skip entries whose path begins with "Retention"?

1 0
replied on October 21, 2019

@████████ - When I removed the LF:Lookin, I got the same result.  THEN, I stripped off all 29 of the .AddColumns except for Id.  I used an EntryInfo object, after that, to get what field values I needed.  It finished in under 5 minutes. 

Does that sound like normal behavior that the # of columns would so greatly impact the performance?

0 0
replied on October 21, 2019

I would say yes. The more columns you add, the more data it needs to join from the repository database and the query becomes increasingly more expensive.

You can see "similar" performance hits when you add too many columns to the client and start browsing folders quickly.

If you don't need the data for the 29 columns, including them is just creating overhead. If you do need the data, it might be more efficient to grab them per document instead of all at once in a massive result set.

2 0
You are not allowed to follow up in this post.

Sign in to reply to this post.

Details
Attachments
Related Posts
Loading ...