none
SQLServer Backup using Veritas Netbackup after Index Rebuild RRS feed

  • Question

  • All,

    This is independent  of the below thread but it gives a context to the issue now at hand.

    We use Veritas Netbackup V7.7.2 to backup the databases onto tape and it has been working fine. But the backup of the databases in the instance is now not working after the Index Rebuild - for none of the databases including the ones which were not subjected to Index Rebuild . We initially thought it is an issue with the Staging environment but it turns out the issue is on Production as well which we tested right after the Index Rebuild completed.

    NF - Error: Failed to query database metadata for 

    However, the restore on the instances are going through fine. The issue is only with Backups. 

    I did a manual SQLServer backup and tried the Netbackup to backup the Transaction Log but the error remains.

    https://social.msdn.microsoft.com/Forums/sqlserver/en-US/a46733fa-7925-425a-a92f-c69db9b04d08/reindex-transaction?forum=sqldatabaseengine

    11/25/2017 10:25:52 - Info nbjm (pid=5324) starting backup job (jobid=8512283) for client nycsqlp1.company.net, policy NYC_WG_SQL_CLOVER_P1_FULL_DAILY, schedule Daily

    11/25/2017 10:25:52 - Info nbjm (pid=5324) requesting MEDIA_SERVER_WITH_ATTRIBUTES resources from RB for backup job (jobid=8512283, request id:{7732BE81-7CB5-4FDA-B4D6-A1652F044A79})

    11/25/2017 10:25:52 - requesting resource WG_DEDUPE_STORAGE_5

    11/25/2017 10:25:52 - requesting resource nycnetbkp1.company.net.NBU_CLIENT.MAXJOBS.nycsqlp1.company.net

    11/25/2017 10:25:52 - requesting resource nycnetbkp1.company.net.NBU_POLICY.MAXJOBS.NYC_WG_SQL_CLOVER_P1_FULL_DAILY

    11/25/2017 10:25:52 - granted resource  nycnetbkp1.company.net.NBU_CLIENT.MAXJOBS.nycsqlp1.company.net

    11/25/2017 10:25:52 - granted resource  nycnetbkp1.company.net.NBU_POLICY.MAXJOBS.NYC_WG_SQL_CLOVER_P1_FULL_DAILY

    11/25/2017 10:25:52 - granted resource  WG_DEDUPE_STORAGE_5

    11/25/2017 10:25:52 - estimated 0 kbytes needed

    11/25/2017 10:25:52 - Info nbjm (pid=5324) started backup (backupid=nycsqlp1.company.net_1511623552) job for client nycsqlp1.company.net, policy NYC_WG_SQL_CLOVER_P1_FULL_DAILY, schedule Daily on storage unit WG_DEDUPE_STORAGE_5

    11/25/2017 10:25:54 - started process bpbrm (pid=7240)

    11/25/2017 10:25:55 - Info bpbrm (pid=7240) nycsqlp1.company.net is the host to backup data from

    11/25/2017 10:25:55 - Info bpbrm (pid=7240) reading file list for client

    11/25/2017 10:25:55 - connecting

    11/25/2017 10:25:56 - Info bpbrm (pid=7240) starting bphdb on client

    11/25/2017 10:25:56 - Info bphdb (pid=37472) Backup started

    11/25/2017 10:25:56 - connected; connect time: 0:00:00

    11/25/2017 10:25:59 - Info dbclient (pid=17884) INF - Error: Failed to query database metadata for <CLOVER>

    11/25/2017 10:25:59 - Info dbclient (pid=17884) INF - Error: Failed to query database metadata for <CLOVER_CLR>

    11/25/2017 10:25:59 - Info dbclient (pid=17884) INF - Error: Failed to query database metadata for <CLOVER_Input_Staging>

    11/25/2017 10:25:59 - Info dbclient (pid=17884) INF - Error: Failed to query database metadata for <CLOVER_Log>

    11/25/2017 10:25:59 - Info dbclient (pid=17884) INF - Error: Failed to query database metadata for <CLOVER_Extension>

    11/25/2017 10:26:01 - Info dbclient (pid=17884) INF - Error: Failed to query database metadata for <DBAArchive>

    11/25/2017 10:26:01 - Info dbclient (pid=17884) INF - Error: Failed to query database metadata for <master>

    11/25/2017 10:26:01 - Info dbclient (pid=17884) INF - Error: Failed to query database metadata for <model>

    11/25/2017 10:26:02 - Info dbclient (pid=17884) INF - Error: Failed to query database metadata for <msdb>

    11/25/2017 10:26:04 - Info dbclient (pid=17884) INF - Results of executing <D:\Program Files\VERITAS\NetBackup\DbExt\MsSql\nycsqlp1-CLOVERT-All-Full-Backup.bch>:

    11/25/2017 10:26:04 - Info dbclient (pid=17884) <0> operations succeeded. <9> operations failed.

    11/25/2017 10:26:04 - Info dbclient (pid=17884) INF - The following object(s) were not backed up successfully.

    11/25/2017 10:26:04 - Info dbclient (pid=17884) INF - CLOVER

    11/25/2017 10:26:04 - Info dbclient (pid=17884) INF - CLOVER_CLR

    11/25/2017 10:26:04 - Info dbclient (pid=17884) INF - CLOVER_Input_Staging

    11/25/2017 10:26:04 - Info dbclient (pid=17884) INF - CLOVER_Log

    11/25/2017 10:26:04 - Info dbclient (pid=17884) INF - CLOVER_Extension

    11/25/2017 10:26:04 - Info dbclient (pid=17884) INF - DBAArchive

    11/25/2017 10:26:04 - Info dbclient (pid=17884) INF - master

    11/25/2017 10:26:04 - Info dbclient (pid=17884) INF - model

    11/25/2017 10:26:04 - Info dbclient (pid=17884) INF - msdb

    11/25/2017 10:26:25 - Error bpbrm (pid=7240) from client nycsqlp1.company.net: ERR - command failed: none of the requested files were backed up (2)

    11/25/2017 10:26:25 - Error bpbrm (pid=7240) from client nycsqlp1.company.net: ERR - bphdb exit status = 2: none of the requested files were backed up

    11/25/2017 10:26:25 - Error bpbrm (pid=7240) cannot send mail to backupsupport@mycompany.com

    11/25/2017 10:26:25 - end writing

    11/25/2017 10:26:26 - Info bphdb (pid=37472) done. status: 2: none of the requested files were backed up

    none of the requested files were backed up  (2)


    • Edited by grajee Saturday, November 25, 2017 5:22 PM Added Transaction Log Backup Info
    Saturday, November 25, 2017 4:17 PM

Answers

  • It turns out the NetBackup code sql code to query the metadata was returning arithmetic overflow error exceeding dataype int.

    Select size * 8/ 1024 from sys.master_files results in error if any of the data files exceed the limits of dataype int.

    • Marked as answer by grajee Wednesday, November 29, 2017 12:46 AM
    Wednesday, November 29, 2017 12:46 AM

All replies

  • The question may be more apt in a forum for Veritas. (A product I know nothing about.)

    At least if you want an answer from an SQL Server perspective, you need find out what this Failed to query database metadata for <CLOVER> means. For instance, you could run Profiler to capture what queries Veritas emits.


    Erland Sommarskog, SQL Server MVP, esquel@sommarskog.se

    Saturday, November 25, 2017 6:03 PM
  • The Tape Backup team is working with Veritas. I as wondring if anyone had offhand experience.
    Saturday, November 25, 2017 10:39 PM
  • It turns out the NetBackup code sql code to query the metadata was returning arithmetic overflow error exceeding dataype int.

    Select size * 8/ 1024 from sys.master_files results in error if any of the data files exceed the limits of dataype int.

    • Marked as answer by grajee Wednesday, November 29, 2017 12:46 AM
    Wednesday, November 29, 2017 12:46 AM
  • A bit embarrassing for the vendor, to put it mildly.

    Erland Sommarskog, SQL Server MVP, esquel@sommarskog.se

    Wednesday, November 29, 2017 10:17 PM
  • Yes. We filed a case on Friday but had to wait till Monday to get a patch. Now that we have the patch we cannot apply it in production during the weekdays because it is peek season for us. I'm doing a sqlserver backup till saturday/sunday and we hope to apply it on saturday itself.
    Thursday, November 30, 2017 12:09 AM
  • A double fail. The app does not report the error correctly, making diagnosis much more difficult. 
    Thursday, November 30, 2017 1:34 AM
  • A double fail. The app does not report the error correctly, making diagnosis much more difficult. 

    LOL.  Yup.  That's why I always fix that kind of bug twice. 

    First fix the reporting, then fix the bug!

    A teachable moment.

    Josh

    Thursday, November 30, 2017 3:12 AM