Forum Discussion

Juergen_Zeiling's avatar
Juergen_Zeiling
Level 3
2 years ago

Autoloader goes offline by drive calibration (Quantum V700, LTO-9)

We changed a LTO-6-Library against a new LTO-9 Autoloder.

A new tape is shown as "unknown media" and the capacity "0 Bytes used" with a green bar that the full tape can be used.

New for us is the experience that each new tape we loaded into the 16-slot-magazine will need a calibration when the tape is "inventoried" the first time. When we start the inventory, the tape will be moved to the picker and the loader shows "calibrating" in the display and in the Onboard-Management console. After ca. 2 minutes the loader goes offline and Backup Exec shows both drive and loader offline. We did make them online again, but the calibration goes further on. After some hour, the drive will load the tape and then we can move the tape by move-command in the management of the loader (we did not try the onboard-management) to the slot, where it comes from. After this we make a "read slot" and the tape will be shown in the slot.

Then we start inventory for this slot again, and the tape will go over the picker to the drive, loaded and unloaded and then the tape return to the slot automatically. The tape is shown in Backup Exec as "empty Medium" and the capacity is shown as "0 Bytes used of 16,4 TB". Now we can use the tape for backup-jobs. No error and no offline will occur.

But whenever we start an inventory for a brand new tape, the drive and the loader goes offline.

We think, this is a time-out-error, because the inventory in the drive continues without error. But the time, how long backup-Exec is waiting for an answer for calibarting is to short, so it make drive and library offline.

Is this correct? Can we adjust manually the parameter for the time-out (Registry?)?
Is there another reason, why the machine goes offline?

In the "bengine_AlwaysOnLog_2022-09-22_10-49-31.log" we can find:
[2556] 2022-09-21T18:40:59.990 [tpfmt] - TF_GetTargetMediaDeviceInfo(): Execute(ADAMM_MOVER_EXECUTE_GET_SERVER_INFO) failed. mvrRet=87.

In the beginning, we did only disable the tape0 (LTO-6), but at the end the device was switched off, removed from SAS-Connector and deleted from Backup Exec. The error consists.

By the way, all drivers are current. The library and the drive are shown correct in device manager.

JuZe

20
windows server 2012

9 Replies

Sort By
  • PJ_Backup's avatar
    PJ_Backup
    Moderator
    2 years ago

    Although you might like to review this rather comprehensive technote on troubleshooting Tape Libraries and Tape Drives:
    https://www.veritas.com/support/en_US/article.100016050 - How to troubleshoot issues with a Robotic Library (autoloader/changer) and/or Tape Drive(s)

    what strikes me straight way is the fact that support for LTO9 didn't come in until at least Backup Exec 21.
    Consequently I would checkout the HCLs for the different versions linkable from here:
    https://www.veritas.com/content/support/en_US/article.100040087 - Backup Exec Compatibility List Home Page

    You need to ensure both LTO9 support is there, as well as making sure the tape library is supported. Sometimes there are useful foornotes in the HCL listing certain caveats. For instance this note sounds applicable in your scenario:

    https://www.veritas.com/content/support/en_US/article.100051049 - Backup Exec and Support for LTO9 tape media.
    (especially the comment regarding timeout when initializing new cartridges)

  • Juergen_Zeiling's avatar
    Juergen_Zeiling
    Level 3
    2 years ago

    Dear PJ_Backup,

    thank you for your qucik answer. We know, BE 20 is not more supported, but the changes to 21 are not so severe concerning the tape or library. When the drivers are OK, the LTO-9-Device should work also in BE20. And the last listet LTO-drive in BE 21 HCL  is LTO-8.

    The Library is an "Autoloader 3", model "V700, which is the same machine we used the years before with an LTO-6-drive. The actual firmware is v100, also the same as before. So this should not be a factor for the error.
    The drive is an LTO-9 from IBM. I got the driver from Quantum-Homepage for Win 2012R2, which is working well with Win 2012.

    The last link from you is exactly describing what happens. I did read about the erasing command, which smells like it is the same reason. But 18 hours are very much. But maybe it is not enough for LTO-9-tapes, which are longer than LTO-8.

    The problem is the timeout by initialising or by erasing the tape, how long BE wil wait untill showing the device offline.

    But my question was and is: is there a command or a registry key, what would exceed in Backup Exec the limit of 16 minutes until the drive and the loader goes offline? And I suspect the parameter is in BE, because the drive itself continues to calibrate, even the drive and library is offline in BE.

    16 minutes is 960 seconds or 0x3C0 in Hex. I cannot find a key in the registry like 960/3C0. Or is this limit hardcoded in BE?

    We did only buy 'not preinitalized tapes', because we did'nt know that there is an offer for this (b.e. called "Spectra Certified Media" in https://spectralogic.com/features/lto-9/). But maybe they are more expensive. And I would prefer to initialize them in my drive. Even I don't think there is a difference. But then I would expect, all the tapes should be preinitialized before the first use.I did hear about a tool, what is initializing the tapes outside from BE. Can you offer a source for that?

    JuZe

  • pkh's avatar
    pkh
    Moderator
    2 years ago

    We changed a LTO-6-Library against a new LTO-9 Autoloder.

    What are the tapes that you are using with the LTO-9 drive?  If they are LTO-6, then they will not work.  LTO drives read tapes 2 generations back and write to tapes 1 generation back.  That is, your LTO-9 drive will read LTO-7 and LTO-8 tapes and write to LTO-8 tapes, but it will not read or write to LTO-6 tapes.

    Also, if LTO-9 is not supported by BE20, then it has not been tested with BE20.  If you encounter problems, you are on your own, no support.

  • PJ_Backup's avatar
    PJ_Backup
    Moderator
    2 years ago

    No idea about the registry entry sorry. There are related RTL keys to be found in:
    HKLM \ SOFTWARE \ Veritas (or Symantec) \ Backup Exec for Windows \ Adamm
    for instance there is an iSCSIWaitTime but I'm just guessing here (on my test system that is set to 20 [hex 14] )

    I did notice from your screenshot that your RTL appears to have the wrong driver. As per that first technote link I gave above, the medium changer should be recognised in Device Manager as "Unknown Medium Changer" and not Quantum SuperLoader. More details in that TN. eg:

    That may or may not make any difference..!

    • Juergen_Zeiling's avatar
      Juergen_Zeiling
      Level 3
      2 years ago

      Dear PJ_Backup,

      thank you for your response.

      It is correct, that the "Unkown Medium Changer" has been changed in our environment to "Quantum Superloader". But we had the same situation before we installed the driver for Autoloader. The autoloader goes offline, when we start initialization in both configurations. There seems to be no difference, if the medium changer has the current driver (from homepage of quantum) or the "unknown" driver.

      I hope, there will be a parmeter to give the autoloader (or any other media changer, like 'scalar') more time before going offline.

      JuZe

      • PJ_Backup's avatar
        PJ_Backup
        Moderator
        2 years ago

        Dear JuZe

        Firstly apologies in advance, you're probably not going to like my reply...

        On the Spectralogic website you bought your tapes from it does warn that:

        "With the release of the LTO-9 tape drive, a new one-time initialization process is required to optimize each new cartridge before it is used for writing data"
        "Although the initialization of LTO-9 media occurs only once, it can be quite time consuming. The process typically takes between 40 and 60 minutes, but can take up to 2 hours. To enable customers to immediately use their LTO-9 cartridges, Spectra Logic provides LTO-9 Certified Media, which is pre-initialized during the certification process."
        " It also means that the host software that is used with the tape library will not experience any delays or timeouts due to the extended initialization time needed on new LTO-9 tapes"

        That taken in conjunction with the footnote for LTO9 tapes in both the BE21 & BE22 HCLs:
        "Failures may occur with 'Initiliaze' jobs (on new tape cartridge) and 'Long Erase' jobs with LTO9 tape cartridge"

        Digging deeper I discovered it is planned (but not guaranteed) for this to be addressed  in BE22.1 - GA date approx: Q1 2023.

        I believe that is why in the technote https://www.veritas.com/content/support/en_US/article.100051049 there is the comment:
        Note:- It is recommended to use already initialized LTO9 cartridges in Backup Exec. This can be done outside of Backup Exec using the vendor provided portal.