![]() |
|
ECGridOS Developer Development related posts. |
![]() |
|
Thread Tools | Display Modes |
#1
|
|||
|
|||
![]()
While running tests of concurrent retrieval of a large set of parcel IDs, ECGridOS seems to occasionally abort some of the retrieval requests with a "Session Timeout" error. Does the network support concurrent logins?
|
#2
|
||||
|
||||
![]()
Multiple sessions should not be a problem.
The problem seems that the Session Timeout was being incremented at the start of the call, not the end of the call. If the call ran longer than the default Timeout (180 seconds), then the session was expired by the time the call returned. I have updated the system to advance the Timeout at the END of the call which should fix this problem. Thanks for finding and reporting this bug! -=tg=- Last edited by Todd Gould; 10-06-2010 at 05:58 PM. Reason: not solved |
#3
|
|||
|
|||
![]()
I don't think that did it. In some of our tests, one of the sessions would get the "Session Timeout" fairly early in the process, and I would probably say it was still within the first 180 seconds. In any case, we just ran another test and the same thing happened.
|
#4
|
||||
|
||||
![]()
I see something else going on. It seems that there is some "unknown" error that is being thrown and the system is reporting a "timeout."
I am researching both the root problem and the false error code and will report back here when I have a solution. -=tg=- |
#5
|
||||
|
||||
![]()
I may have found the culprit. In each case it seems you were attempting to download the same file at the same time with two sessions.
There is nothing specifically wrong with that, but the code had not anticipated that and had implemented a read-lock on the file. When the second session tried to read the same file, it was passed an Access Denied message from the OS. Then it gets worse, we were not properly tracking the error and it ended up being wrongly reported as a Timeout error (in a strange way, that wasn't entirely untrue). I will have a patch for this implemented and on-line tonight. Note that this update will only be for the v2.2 version of ECGridOS. -=tg=- |
#6
|
||||
|
||||
![]()
There may be another bug in there. I'm still looking and will flush them all out.
-=tg=- |
![]() |
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
Display Modes | |
|
|