SS 3.0 Stop Scraping Session not working
Hi:
I'm trying to stop / kill / abort a scraping session that's been running for the past hour or so. The "Stop" button changed to "Start" after I clicked it, but session continued. The right-click and select "Stop Scraping Session" on the session branch doesn't stop it either -- the log continues to roll, the CPU is still 50% (on a two CPU system).
I'm running v3.0, with the sessio invoked from the workbench (not from a command line driver)
I'm reluctant to just kill SS, but there doesn't seem to be any other way to stop this runaway session.
Retested this problem, tried to kill session shortly into the restart. Here's what the end of the session log looks like:
<... snip ...>
Stopping the scraping session.
Stopping the scraping session.
Stopping the scraping session.
Stopping the scraping session.
Stopping the scraping session.
Stopping the scraping session.
Stopping the scraping session.
Stopping the scraping session.
Scraping session finished.
------------------------------------
This session is run from the workbench and is "driven" by a script at the start that reads in file names to feed the extraction pattern. Evidently "Stop" only stops the current instance of the scrape, not the session itself. So the session continued to read in and feed the names from the file to the extractor and then kill the scrape. Since there are about 12,500 names in the file, this feed/kill takes a while.
It doesn't appear to me that this is the way the developers intended "Stop" to work. It is a button on the Session General tab, and so should kill the driver script that hangs from that level as well, it seems to me.
-- Roy Zider
SS 3.0 Stop Scraping Session not working
Thanks, Scott.
I do of course have smaller test sets -- "special", "special2", "8", "200" and so forth to build up to the big one, and do regression testing against prior problem cases.
I was confused at first as to what was happening, since I didn't have autoscroll on and the reading messages were accumulating down below the visible portion of the session log.
Thanks for you clarification.
-- Roy
Runaway scrapes when reading from a file
Roy,
This is one aspect of screen-scraper that is dependent on the framework on which it is built. As the lead developer of screen-scraper has described to me when you tell screen-scraper to stop scraping screen-scraper sends a request to the Java BeanShell intermediary. It is then up to BeanShell to determine when it is safe to execute the command sent to it. This helps protect the Java Virtual Machine (JVM) as it's executing on-the-fly.
In this case, the reason BeanShell is not immediately executing the command you sent it is because it is reading from a file and is ensuring that by closing out the file reading process it does not corrupt that file on your system.
While you're testing I would recommend you read from a file with, say, 5-10 names rather than the 12,500. Once you're through that phase of testing run it through the entire list and hopefully it will parse the file and not hang.
You were wise not to kill screen-scraper from the Task Manager as this can some times corrupt the screen-scraper database.
Thank you,
Scott Wilson