INFO: IIS: Users See Error Message: The Remote Procedure Call Failed and Did Not Execute

This article describes when IIS displays the following error message in a browser:

This error is displayed in the Web browser when a request is made by a user to an "out-of-process" Web application that has terminated or has been killed. This is a temporary message that is displayed until IIS has restarted the failed Web application. IIS, by default, automatically restarts all failed out-of-process Web applications, and this behavior is governed by a metabase value, "AppOopRecoverLimit".

In IIS 4.0, a faulty in-process Web application can cause the entire Inetinfo.exe service to fail, which renders all Web sites unavailable. To help alleviate this situation, IIS 4.0 introduced the concept of out-of-process Web applications. When a Web application is designated as out-of-process, the application runs in a separate process (MTX.exe) that is isolated from the main IIS service (Inetinfo.exe) so that if the faulty Web application crashes, the main IIS service is unaffected and all other in-process and out-of-process Web sites remain available.

In IIS 4.0, the AppOopRecoverLimit is not present in the metabase by default, but the value internally defaults to "5" for all out-of-process Web applications. This means that you can kill/crash an out-of-process Web site up to five times and IIS will restart the Web application for you automatically. In IIS 4.0, this setting only applies to out-of-process Web applications that run in MTX.exe.

In IIS 5.0, the process isolation model was changed to encompass the new medium isolation "pooled out-of-process" Web applications in addition to high-isolation out-of-process Web applications. In IIS 5.0, the AppOopRecoverLimit limit is not present in the metabase by default, but the value internally defaults to "-1", which means IIS will restart all medium-isolation and high-isolation Web applications indefinitely. If the AppOopRecoverLimit is set in the metabase on a medium-isolation (pooled) Web application, it will be ignored and the pooled out-of-process Web application will restart indefinitely by default. In IIS 5.0, this setting can only be used on high-isolation Web applications that are running in DLLHost.exe. Additionally, if the Web site is in low isolation (in-process), then the application restart capability is governed by the properties for the IIS Admin service, which are listed in the Service Control Manager (which is configured through the "Services" Microsoft Management Console [MMC] snap-in, located in the Administrative tools group) and not by the AppOopRecoverLimit metabase entry. In IIS 5.0, by default, if the Inetinfo.exe service crashes, the IISReset.exe command will be issued to restart IIS indefinitely.

Each time IIS 4.0 or 5.0 has to restart a failed out-of-process Web application, an error message similar to the following is logged in the event log: In addition, each time that an out-of-process IIS Web site crashes or is killed, the next user to request a page from that Web site will receive the following error message once in his or her browser while IIS is restarting the failed site: After receiving that error message, if the user requests the page again (or clicks refresh in the browser), IIS will have restarted the failed out-of-process Web site (provided that the AppOOPRecoverLimit hasn't been reached) and will return the page to the browser as usual.

In IIS 4.0, after the out-of-process Web site has failed five times, on the sixth attempt to restart that Web site, IIS will not restart the site, and it will instead return the following error message to the browser: Then, if the user clicks refresh or requests the page again, he or she will then receive the following error message: If the user continues to click refresh in the browser or to request a page from the failed out-of-process application, he or she will continue to see the "Server Application Error" and the Web application will remain stopped until IIS is restarted.

In addition, in the IIS logs if you are using W3C Extended Logging, you may see an entry like the following for each GET request that is sent to the server once AppOOPRecoverLimit has been reached: Note in the third column from the right the "502" response code from IIS and the column next to that containing "1727" and "1726", which are the "Win32" response codes. A 1727 code maps to the "The Remote Procedure Call Failed and Did Not Execute" error and 1726 maps to "The Server has reached the maximum recovery...." The "Win32 Status" field is not enabled by default. In order to make that column show up in the IIS logs, you must edit the extended logging properties for your Web site.

Also note that the time recorded in the IIS logs is in Greenwich mean time. So, for example, if you are in the (U.S.) eastern time zone, you would have to subtract 5 hours from the time that is logged in the first column to determine the local time at which the request was received (eastern time is Greenwich mean time - 5).

For additional information about some sample codes, click the article number below to view the article in the Microsoft Knowledge Base: For additional information about setting the AppOopRecoveryLimit, click the article number below to view the article in the Microsoft Knowledge Base: