We are currently actively working on a fix to overcome problems with temporarily unavailable data in MARS.

If you get the MARS error 'Requested data is on one or more damaged tape: tape_name' please use the form below to check for the relevant tape_name. Below you can also search a list of common ECMWF datasets affected by these tape problems. The search results will give you a good indication on the data currently not accessible, e.g. the line

/mars/prod/tigge/o/g/cf/sfc/marstiprod/glob/cf/sfc/2018/20181023.0600.222067.20181025.053439

shows that the Surface (sfc) Control Forecast (cf) TIGGE global production data for the 06 UTC run on 23 Oct 2018 is currently not available.

If the data you are interested in is listed in any of the search results we are already aware of the issue and working on the data recovery. There is no need to notify us about the problem. Please check here again later. For any other problem with MARS access, please visit the online documentation.

Feel free to Ask a question about damaged tapes

We apologise for the inconvenience any temporary unavailability of data may cause you.


36 Comments

  1. Bruce Ingleby

    I can't verify against operational analyses for 2022 because tape J1007400 is damaged (it contains months of surface fields from 0001). Any indication at to if/when this will be fixed would be useful.

  2. Paul Poli

    Hello, thanks for exposing the issues – much appreciated to know that this is being working on already. I was trying to access ERA5 OFB data from a tape that MARS reports as damaged, and it is listed here already (J1495400). This tape contains observation feedback data that we would need to access, in order to prepare for ERA6. Does one know (even if, very approximately) how long the data recovery make take?

  3. Dominique Lucas

    Hello Bruce and Paul,

    A tape can be damaged in various ways and its content may be recovered directly by us in just a few days or we may have to send the damaged tape to the manufacturer for remote recovery, which may take up to several weeks or even longer. The good news for you two is that the content of tape J10007400 has already been recovered and that we will be able to recover the content of tape J1495400 ourselves. This shouldn't take too long.

    REgards,

    Dominique

  4. Jasper Denissen

    Hey all,

    For an offline surface experiment, I want to use meteorological reanalysis forcing from ERA5, which seems to be unavailable starting 20201108 (AccessError: Requested data is on one or more damaged tape: J1911100). Any indication when this issue could be resolved?

    Cheers,

    Jasper

    1. Dominique Lucas

      Hello Jasper,

      We should also be able to recover ourselves the data on tape J1911100, This should happen shortly.

      Regards,
      Dominique

  5. Michael Rennie

    Hi,

    I just encountered problems with tape J1051000 (operational analysis in 2019, using it in iver). Is this likely to resolved any time soon?

    Kind regards,

    Mike

    1. Dominique Lucas

      Hi Mike,

      Tape  J1051000 is being processed at ECMWF. This shouldn't take too long.

      Dominique

  6. Ernest Koffi

    Hello, I have experienced the error message below for all my experiments on the same day. It is about an issue related to a damage tape when running the task fetchmars:  etest -f ec:/emos_backup/ELDA/0001/2021020712/wavelet_out_T159_L137_2021020712.cv
    6:56:22 fetchmars[1205] echo 'Wavelet JB: ,' /ec/fws2/sb/work/rd/cxek/i0fy/eda_err/JB/wavelet_out_T159_L137_2021020712.cv , 'from ,' ec:/emos_backup/ELDA/0001/2021020712/wavelet_out_T159_L137_2021020712.cv
    Wavelet JB: , /ec/fws2/sb/work/rd/cxek/i0fy/eda_err/JB/wavelet_out_T159_L137_2021020712.cv , from , ec:/emos_backup/ELDA/0001/2021020712/wavelet_out_T159_L137_2021020712.cv
    6:56:22 fetchmars[1206] [[ -f xxxx ]]
    6:56:22 fetchmars[1210] [[ ! -f /ec/fws2/sb/work/rd/cxek/i0fy/eda_err/JB/wavelet_out_T159_L137_2021020712.cv ]]
    6:56:22 fetchmars[1211] rm -f xxxx
    6:56:22 fetchmars[1212] etest -f ec:/emos/ELDA/0001/2021020712/wavelet_out_T159_L137_2021020712.cv
    6:56:23 fetchmars[1212] ecp -o ec:/emos/ELDA/0001/2021020712/wavelet_out_T159_L137_2021020712.cv xxxx
    ecp: copying from ec:/emos/ELDA/0001/2021020712/wavelet_out_T159_L137_2021020712.cv: Data offline; requested file /emos/ELDA/0001/2021020712/wavelet_out_T159_L137_2021020712.cv is on damaged tape volume J1449000
    06:56:23 slurm_script[1] ERROR

    Ernest 

    1. Dominique Lucas

      Hello Ernest,
      This ECFS file is available again. Note that this page only reports on MARS data unavailabilities.
      Dominique

      1. Ernest Koffi

        Thanks! It is working fine

        Ernest 

  7. Marcus Koehler

    Hello, I'm experiencing difficulties with tape J1891000.  I had retrieved data from there a few weeks ago if I'm not mistaken.  Has anything happened to the tape and is there an approximate date by when this could be fixed again?

    Many thanks!

    Marcus

    mars - ERROR  - 20230725.133036 - AccessError: Requested data is on one or more damaged tape: J1891000. For more information, visit MARS data unavailability in ECMWF tape library [marser]
    mars - ERROR  - 20230725.133036 - Error code is -2

    1. Dominique Lucas

      Hello Marcus,

      Tape J1891000 is indeed currently not access, but its content is being repacked and should hopefully be available again soon.

      Dominique

      1. Marcus Koehler

        Great, thanks for letting me know!  Marcus

  8. Christopher Roberts

    Hi, the issue with tape J3324000 is stopping some aspects of high-resolution evaluation of cycle 49R1. Is there any estimate of when this data will be available again? Thanks1  

  9. Jasper Denissen

    Hi,

    I have an issue with tape J0035900:

    ecp: copying from ec:/eras/era5l/OSM_FORCING/forcing_ea_oper_1_19900220.tar.gz: Data offline; requested file /eras/era5l/OSM_FORCING/forcing_ea_oper_1_19900220.tar.gz is on damaged tape volume J0035900

    Any comment on when it would be available again?

    Thanks!

    1. Dominique Lucas

      Hello Jasper,

      We are trying to recover the content of tape J0035900  at ECMWF. This should not take more than a few days. If this doesn't succeed, we'll have to send the tape to a remote site, which will take longer.

      KInd regards,

      Dominique

      1. Jasper Denissen

        Hey Dominique,

        I tried again this morning and it seems the issue is resolved.

        Thank you.

        1. Dominique Lucas

          Hey Jasper,

          Good news! Thanks for confirming this.

          Dominique

  10. Marcin Chrust

    Hi,


    My experiment i50d on tape J3080800 is affected by a 'damaged tape' issue. Is there hope for a quick recovery of this tape?

    Thanks,
    Marcin

    1. Paul Dando

      H Marcin,

      Tape J3080800 is in the process of being repacked.  All being well, the data should be available again later today or tomorrow.

      Paul


      1. Marcin Chrust

        Hi Paul,

        Great. Thanks for letting me know.

        Marcin

        1. Paul Dando

          Good morning Marcin,

          I'm told the tape has been repacked successfully so the data for your i50d experiment should be available now.

          Paul

  11. Annelize Van Niekerk

    Hello,

    A couple of my experiments have been archived onto damaged tapes, they are i4tc and i4qd.

    The error says they are on tapes J3080800 and J3398100, the latter having been out of action for quite some time now.

    Is there hope to recover them shortly because they are quite costly experiments to run.


    Thanks very much,


    Annelize

    1. Annelize Van Niekerk

      just seen that J3080800 shouwl be back now.

      1. Annelize Van Niekerk

        and the other seems to be working too!

  12. Ernest Koffi

    Hello, 

    I have two experiments that are failing with the following message. Any idea? 

    Experiment i4ce failed at ecflow-pifs-rd-k-1:/cxek/i4ce/an/main/da00/an/compo_bfas on aa6-188 JobID 1835800\nOutput is in /ec/fws2/sb/log/rd/cxek/i4ce/an/main/da00/an/compo_bfas.2\n\nLogfile:\n
    mars - ERROR  - 20230927.093919 - Mars server task finished in error
    mars - ERROR  - 20230927.093919 - AccessError: The tape on which the data reside is unavailable. Please visit ECMWF Support Portal and raise a ticket: http://support.ecmwf.int [marsod]
    mars - ERROR  - 20230927.093919 - Error code is -2
    mars - ERROR  - 20230927.093919 - Last error is -2
    mars - ERROR  - 20230927.093919 - Expected 18, got 0.
    mars - ERROR  - 20230927.093919 - Request failed
    mars - INFO   - 20230927.093919 - Memory used: 39.76 Mbyte(s)
    mars - ERROR  - 20230927.093919 - Some errors reported (last error -1)

    1. Paul Dando

      Hello Ernest,

      For this one, please can you follow the instructions and visit the ECMWF Support Portal and raise a ticket: http://support.ecmwf.int.  Please include in the ticket  the MARS retrieval than generates the error.  There is no tape volume mentioned so we need to make a record of this and investigate.

      Thanks

      Paul


  13. Ernest Koffi

    Thank Paul! Yes, I open the ticket as you suggested above. 

  14. Ernest Koffi

    Hello,

    I'm experiencing damage tape for two of analysis experiments with a failure in fetchmars taks as follows:

    0:27:26 fetchmars[1212] ecp -o ec:/emos/ELDA/0001/2022120900/wavelet_out_T191_L137_2022120900.cv xxxx

    ecp: copying from ec:/emos/ELDA/0001/2022120900/wavelet_out_T191_L137_2022120900.cv: Data offline; requested file /emos/ELDA/0001/2022120900/wavelet_out_T191_L137_2022120900.cv is on damaged tape volume K0452600


  15. Dominique Lucas

    Hello Ernest,

    Could raise another ticket for these missing files in ECFS? We are currently recovering them from a backup.

    Thanks,
    Dominique

  16. Ernest Koffi

    Hello Dominique, 

    I would like to, but I do not have the link for ECFS. 

    Ernest 

    1. Dominique Lucas

      I see. You will also use our support portal (http://support.ecmwf.int/) to create a new ticket for ECFS. For example, 'Report a problem on computing' under https://confluence.ecmwf.int/category/sc will do.

      Thanks.

      Dominique

      1. Ernest Koffi

        Thanks Dominique. I have just created the ticket. 

        Ernest

  17. Patrick Laloyaux

    Hello, I hope this is the right place to post. I got this from MARS

    AccessError: Requested data is on one or more damaged tape: K0534800. For more information, visit MARS data unavailability in ECMWF tape library [marser]

    Thank you to anyone who can help on this matter

    1. Patrick Laloyaux

      Is this page monitored? Any update? Shall I report somewhere else?

      1. Dominique Lucas

        Hello Patrick,

        I can confirm that the tape you mention is not available. This interactive wiki page allows you to enquiry about any tape containing MARS data and reported as 'damaged'. The only additional information I have is that tape K0534800 is currently not in the tape library. To get further details about this tape and its recovery process, could I invite you to report an Jira issue under our support portal https://confluence.ecmwf.int/site/support ?

        Thank you.