You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
FhirClient incorrectly interprets FhirClientSettings.Timeout milliseconds as seconds, when HTTP Timeout value is set.
The defect is in the HttpClientRequester constructor, which incorrectly uses the 4-param TimeSpan constructor (public TimeSpan (int days, int hours, int minutes, int seconds)). FhirClientSettings.Timeout is in milliseconds (as documented, and as indicated by the default value of 100000), but is passed into the seconds param. Presumably the intent was to use the 5-param constructor, of which the 5th param is milliseconds.
This defect can result in incredibly long (e.g., 16+ hour when Timeout is set to 60s) hang times when the network drops, as HttpClient gets stuck waiting to timeout. Temporary network outages cause processing to entirely halt (presumably for 16 hours 40 minutes in our case, though we have never actually waited that long).
Version used:
FHIR Version: STU3
Version: 3.8.3
Clearly, this is a very old version (we are currently working through some issues to get upgraded, but will be on 3.x for the near future). But it seems severe/simple enough that a fix might be of interest. At the very least, documenting seems useful.
The text was updated successfully, but these errors were encountered:
FhirClient incorrectly interprets FhirClientSettings.Timeout milliseconds as seconds, when HTTP Timeout value is set.
The defect is in the
HttpClientRequester
constructor, which incorrectly uses the 4-paramTimeSpan
constructor (public TimeSpan (int days, int hours, int minutes, int seconds)
).FhirClientSettings.Timeout
is in milliseconds (as documented, and as indicated by the default value of 100000), but is passed into the seconds param. Presumably the intent was to use the 5-param constructor, of which the 5th param is milliseconds.This defect can result in incredibly long (e.g., 16+ hour when Timeout is set to 60s) hang times when the network drops, as HttpClient gets stuck waiting to timeout. Temporary network outages cause processing to entirely halt (presumably for 16 hours 40 minutes in our case, though we have never actually waited that long).
Version used:
Clearly, this is a very old version (we are currently working through some issues to get upgraded, but will be on 3.x for the near future). But it seems severe/simple enough that a fix might be of interest. At the very least, documenting seems useful.
The text was updated successfully, but these errors were encountered: