Unconfigured Ad Widget

Collapse

Announcement

Collapse
No announcement yet.

Saving the IFS - What's going on?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Saving the IFS - What's going on?

    F-Rod wrote: > How can I tell what is going on between midnight and 06:30? I'm not sure, but I'd recommend checking your Sav command and excluding the QNTC system. Then check your results. Bill

  • #2
    Saving the IFS - What's going on?

    There are 0 objects in QNTC.

    Comment


    • #3
      Saving the IFS - What's going on?

      Change the logging level of the JOBD for the job that runs the backup. It will put more information in the joblog. Also, you need to exclude QNTC in the sav command. Even if it backed up 0 objects, it could have taken hours to look through each network share.

      Comment


      • #4
        Saving the IFS - What's going on?

        I see. I will try that suggestion because it makes sense, but I can't see how that could have changed in the past few weeks. I mean, was QNTC originally omitted and is now included? Difficult to find out, but I'm really curious as to what may have changed. Thanks.

        Comment


        • #5
          Saving the IFS - What's going on?

          BTW, the jobd is set to 4 00 *SECLVL

          Comment


          • #6
            Saving the IFS - What's going on?

            Our IFS save is also unreasonably slow. In restricted state it takes over 3 hours to save less than 50GB of IFS files. Our SavLib *NONSYS saves over 20GB in less than 15 minutes. We changed it to omit QNTC - no improvement. And there are no entries in the joblog between when the SAV of IFS starts and when it ends over three hours later, even though it's 4 00 *SECLVL.

            Comment


            • #7
              Saving the IFS - What's going on?

              How much of your IFS data is stored in QDLS? QDLS is the slowest of all IFS systems, and should be used only when required. Dave

              Comment


              • #8
                Saving the IFS - What's going on?

                QDLS should be saved with SAVDLO. It's usual to exclude it from the SAV of the IFS. Regarding F-Rod's problem, we would occasionally run into SAVs that took longer than normal. There didn't appear to be any reason for it, and they would eventually return to normal. I never had the time to have IBM really dig into the problem.

                Comment


                • #9
                  Saving the IFS - What's going on?

                  Are you excluding QSYS.LIB ? If not you are backing up everything native all over again. Dave

                  Comment


                  • #10
                    Saving the IFS - What's going on?

                    We do omit QDLS and QSYS.LIB. Below is our SAV. SAV DEV('/QSYS.LIB/TAP01.DEVD') OBJ( ('/*') ('/QSYS.LIB' *OMIT) ('/QDLS' *OMIT) ('/QOPT' *OMIT) ('/QNTC' *OMIT) ) Output(*Print) InfType(*Err) ENDOPT(*Unload) UpdHst(*Sys) IBM has us doing several things that MIGHT fix it, including applying the latest Cume package, but I'm not very hopeful. They also said saving small IFS files is slower than saving larger files, and I tend to believe it. But I don't like it since our files are considered small (36K average). It doesn't seem reasonable to take over 3 hours to save less than 50GB with a new LTO-3 tape drive and a new System i. Geno Geno

                    Comment


                    • #11
                      Saving the IFS - What's going on?

                      Our BRMS SAV save when done on a Sunday night starts reaching out to all the other Hosts in the network, taking about 5 minutes each to determine that these PC's are not powered up, and I often cancel this long running job. ( We see these messages in the joblog as it runs. ) But when saved on Saturday night, the SAV does not look at all these and runs in the normal timeframe. I suspect a WRKLNK - Network Share to some other area not active on Saturday night, but never bothered to track it down. Thanks, I will definitely have to check out the QNTC area for this.

                      Comment


                      • #12
                        Saving the IFS - What's going on?

                        We have had our fair share of IFS saving issues. We had one server that never would complete a system save because of the IFS. During that debugging process, IBM stated that having a lot of small files really slows down the process. Also, if you are on version 1 of the IFS and not version 2, all processing will be slower. If I remember correctly, V2 came out with V5R2 or V5R3 but you had to run a conversion process to get it. You can apply some PTF's to make it available on V5R1. I do have to ask why you are performing a system save while not in restricted state?

                        Comment


                        • #13
                          Saving the IFS - What's going on?

                          I'm looking into a longer-than-normal running backup issue: a full system save via CL. System is not restricted, but I'm not getting any "object in use" msg for the libraries, all are being saved fully except for QUSRSYS and some other lib. The IFS is being saved with options SAVACT(*SYNC) and SAVOPT(*ALWCKPWRT) The SAVLIB portion is done by midnight and the SAV command starts running at that time. Problem: There are no entries in the job log or QHST until after 06:30 when it starts to show the objects not saved; only 15 or so objects are not saved. How can I tell what is going on between midnight and 06:30?

                          Comment


                          • #14
                            Saving the IFS - What's going on?

                            For help troubleshooting this, I would change your InfTyp parm in the SAV command from "*Err" to "*ALL". This will generate details of every object passed over, whether saved or not, including time stamps, I think. Hope this helps to pinpoint the culprit. -Rob

                            Comment

                            Working...
                            X