1 2 3 Previous Next 32 Replies Latest reply on Apr 22, 2016 3:05 PM by Rob Slattery Go to original post
      • 15. Re: patch deployment - reboot

        Look at this:

         

        https://community.shavlik.com/docs/DOC-23833

         

        section: Resolution

         

        Why BSA doesn't perform reboot in such exit code of one of patching task?

        When such reboot can occurs? (simple question) 

         

        BR

        RD

        • 16. Re: patch deployment - reboot
          Bill Robinson

          that's what the exit code override setting is for in the patch global configuration.

          • 17. Re: patch deployment - reboot
            Jim Wilson

            Robert, the bltargetjobmager process is the one that hibernates during the end of job reboot, and reawakens at startup to check the status and inform the application server

             

            We need to see the bltargetjobmanager log to see what happened.

             

            The Job Run Log screen shot above and the bldelpoy previously attached log both show that reboot was requested by BSA.  And the bldeploy log shows a 3 minute gap from when reboot was requested, until the next message

             

            Please check the Windows System Event Log for that time

             

            05/06/15 13:34:53.346 INFObldeploy - Attempting shutdown REBOOT=true TIMEOUT=0 MSG=System is rebooting
            05/06/15 13:37:19.992 INFObldeploy - Bldeploy started for package: 50f4064e5df235b8bc26a87739fc3b8c
            • 18. Re: patch deployment - reboot

              Unfortunately I reverted snapshot for that box so I need to wait till deploy job will finish.

              bltargetjobmanager in debug mode contain 6MB data but I'll send it.

              • 19. Re: patch deployment - reboot
                Bill Robinson

                Jim – one of the patches failed w/ an unhandled exit code.  the error was that the patch was already installed, but blade was not setup to handle this.  so the job failed due to that. iirc if the deploy fails no reboot happens.

                 

                He needs to update the exit code mapping for the ‘already installed’ exit code.

                • 20. Re: patch deployment - reboot

                  It would be nice if BMC could provide such exit codes for windows boxes (in Patch Global Configuration).

                  When I as BSA user may realize that such code should be added in Patch Global Configuration during Change Window?

                  As far as I remember only 3010 exit code was provided.

                  • 21. Re: patch deployment - reboot
                    Jim Wilson

                    Bill, I usually believe you unreservedly, but I remain to be convinced about this one :-)

                     

                    In this case I see in the bldeloy log the message indicating that a reboot was requested, and a 3 minute gap, then the resumption to do the clean up of the staging area.

                     

                    What happened in that 3 minutes?

                    Did the server reboot?

                    Did BSA properly submit the reboot request to the system, but the reboot did not occur?

                     

                    IMHO, only the bltargetjobmanager log and the Windows System Event Log will tell us that...

                     

                    Robert Dołęga please provide the data to clarify whether the target rebooted or not.

                    • 22. Re: patch deployment - reboot
                      Bill Robinson

                      yes - that's true:

                      05/06/15 13:34:53.346 INFO bldeploy - Attempting shutdown REBOOT=true TIMEOUT=0 MSG=System is rebooting
                      05/06/15 13:37:19.992 INFO bldeploy - Bldeploy started for package: 50f4064e5df235b8bc26a87739fc3b8c

                       

                      but this error at the end:

                      05/06/15 13:37:31.442 INFO

                      bldeploy - Bldeploy done - nRet = 11 (Apply partially successful with failed items, overall job phase failure) exitCode = -4001 (Deployment failed)

                       

                      which caused the job to be tagged as failed is likely due to these return codes:

                      Line 4025: 05/06/15 12:46:06.828 ERRORbldeploy - [22][Windows6.1-2008-R2-SP1-KB2478662-x64.msu-MS11-039-en-WINDOWS SERVER 2008 R2 DATACENTER (X64)-SP1] Command returned non-zero exit code: -2145124329

                       

                      05/06/15 12:56:01.781 ERRORbldeploy - [42][Windows6.1-2008R2-KB2705219-v2-x64.msu-MS12-054-en-WINDOWS SERVER 2008 R2 DATACENTER (X64)-SP1] Command returned non-zero exit code: 2359302

                       

                      (missed that first one).  so the 1st one should indicate an actual error w/ the install.  the 2nd one as mentioned needs to be put into the override.

                       

                      so whether or not the box rebooted, i think the failure of the job is related to the two patches that failed to install.

                      • 23. Re: patch deployment - reboot

                        When I've added: 2359302 and -214512432 in Patch Global Configuration, job finished with exit code zero. So Bill, you were right it's matter of handling errors, but above codes (and I gues not only) should be provided by BMC.

                         

                         

                        BR

                        RD

                        • 24. Re: patch deployment - reboot
                          Bill Robinson

                          214512432 is an actually error though isn’t it ?

                           

                          The exit codes are documented by Microsoft on that page i posted.  we don’t want to ignore all of them because many of them are actually errors.  We’d have to look through and see what others are possibly ok to handle/ignore.

                          • 25. Re: patch deployment - reboot

                            You are right, especially that the list is rather static. And some of them should return error, some warning, some reboot etc. It's my suggestion.

                             

                            BR

                            RD

                            • 26. Re: patch deployment - reboot
                              Bill Robinson

                              Yeah – i’ll look into which ones would make sense to default to the various exit code handling.  You can still create an idea to track though ☺

                              • 27. Re: patch deployment - reboot
                                Steve Cupp

                                Robert. . . .just to confirm, what return codes location in the Patch Global Configuration did you put "2359302 and -214512432 in Patch Global Configuration" ? I am currently getting several patches that fail with the -214512432 code, which according to Microsoft means: "install is not needed because no updates are applicable". If I am understanding this correctly, by inserting this code in the proper returns code location, the patch job will not show a failure if this code is returned.

                                • 28. Re: patch deployment - reboot

                                  -214512432  indicates that someone patch manually win box after you created patch catalog.

                                  And this number you shoul put in return codes success - i don't remember the proper label.

                                  Because if in your deploy job you don't have "ignore reboot" this exit code reboots your box unless you add -214512432 to success return codes.

                                  • 29. Re: patch deployment - reboot
                                    Yanick Girouard

                                    I have suggested this as part of standard practices several times (adding this kind of code to either the success return codes or reboot return code list). This should be done by default when you install the application IMO...

                                     

                                    We have added these codes to the success codes for this reason:

                                     

                                    -2145124329,2359302

                                     

                                    One is that the patch is not applicable, and the other that it's already installed. Both may occur when the Shavlik engine detects a lot of missing patches, but that some are actually explicitely installed by one of the patches in the same set, or that the system requires a reboot before a required patch is applied completely to allow the installation of the latter one (still in the same set). It's normal that Windows requires multiple reboots to fully apply a large set of patches.