bulk "force re-check" fails by attempting to run all check simultaneously

[brainchild0]

Generally, the "Force Re-check" operation applied in bulk to a large set of active torrent targets succeeds, by placing most of the items into a queue. Such is the behavior I have generally seen, including in my active deployment, each of which comprises a GTK thin client connected to a remote daemon.

Recently, I attempted to recheck all torrents in a set of about one hundred. The operation failed for most of the torrents, not because the data was found corrupt, but because the application attempted to start all checks simultaneously. After the attempted operation, most items in the set showed on the list as errors ("red water drop icon"), and on the status page with the message "Error: To many open files".

The workaround was to select smaller segments from the list manually, and to wait for the operation to near completion for each segment before moving to the next.

I have found no reason why the problem might have emerged. I have a comparable deployment on another site, with the same client and server configuration, and the problem is not occurring.

Both server and client are version 2.0.3 running on Linux Mint systems.