Summary: | [ftp] FTP fails to re-connect automatically due to 30 sec timeout cache | ||
---|---|---|---|
Product: | [Tools] Target Management | Reporter: | Samuel Wu <samuelwu> |
Component: | RSE | Assignee: | dsdp.tm.rse-inbox <tm.rse-inbox> |
Status: | NEW --- | QA Contact: | Martin Oberhuber <mober.at+eclipse> |
Severity: | normal | ||
Priority: | P3 | Keywords: | helpwanted |
Version: | unspecified | ||
Target Milestone: | --- | ||
Hardware: | PC | ||
OS: | Windows XP | ||
Whiteboard: |
Description
Samuel Wu
2011-05-19 15:21:28 EDT
I've updated the summary, previous value was: FTP connection failed to connect for 30 seconds Resetting the timer after an exception seems like a good idea at first sight ... my concern would be that if the server went offline completely, we might end up in a loop trying to reconnect forever? The 30 sec timeout was introduced in order to avoid constantly probing with NOOP. By far most of the time, FTP connections are stable ... I'd hate compromising the 99% case just for a 1% failure case, so if we change something here we should be very sure it's the right thing to do. Perhaps the simplest possible change would be reducing the timeout to 5 sec. Probing by NOOP every 5 sec seems acceptable to me on a healthy connection, and may also be acceptable in the failure case. I won't have time to investigate this any further, but I can review patches and provide advice. From what I noticed, the RSE FTP client was in a kind of passive mode. It only tried to connect to the host when the user did some action such as save the file, refresh the directory etc. It should try to connect to the server any way in this case. The risk of falling in the loop is not high. To decrease the threshold to 5 seconds sounds fine. |