Forum Discussion

Cupis's avatar
Cupis
Level 4
9 years ago

Retry Jobs and Media Erase (Failed Jobs)

Hi Everyone,

There are two parts to my post, but all relating to a failed job.

Retry Jobs:

In some other products, if your back were to fail and you select to retry the job, it will not backup the entire directory from scratch again, but rather just the items that failed. This would greatly help us, for jobs that are large and take hours to complete, rather than rerunning the entire job just to cover a few files. We sometimes do this manually now by setting up a second job and including only the files that were missed. Is there something that can do this? I understand that it my catch some files that were backed up the first time since there may have been some changes on the files with the snapshot, but I am hoping it would still dramatically shorten the job time.

 

Erasing Jobs on Media:

This one I can guess is not possible, but I am hoping maybe it is anyway. Is there a way to only erase a job from the media. For example I have a tape with three jobs on it, but the 3rd actually failed and we want to erase the failed job from the tape to free up space? Is it possible to do the erase only on that job without erasing the whole tape?

 

Backup Best Practices:

So we want to try reduce the amount of failing jobs we get, which in the grand scheme of things is not alot, but we really want to cut it down as much as possible. Are there any best practices about how big backup jobs should be? For example is it advisable that we have backups that are up to 10Tb? Or should these sort of things be broken down to smaller jobs? not sure if there are any Best Practices involving this.

 

Kind Regards,

Stuart

 

 

  • BE 15 does have checkpoint/restart which allows you to restart jobs.

    You cannot delete specific backup sets from a tape.  The only way to erase a tape is to overwrite it, but this will erase everything on the tape.

    There is no guideline as to how big your job should be.  It is up to the installation to decide.  Having everything in one big job means less scheduling problems, but increases the risk of job failures.

5 Replies

  • BE 15 does have checkpoint/restart which allows you to restart jobs.

    You cannot delete specific backup sets from a tape.  The only way to erase a tape is to overwrite it, but this will erase everything on the tape.

    There is no guideline as to how big your job should be.  It is up to the installation to decide.  Having everything in one big job means less scheduling problems, but increases the risk of job failures.

  • Hi Pkh,

    Thanks very nice feature, just reading through it now. 

    One question around that though. If a job fails because it has encountered a "corrupt file" the restart is not going to help anything as it will use the same snapshot from the beginning to resume? Or does it take a new snapshot at that point to resume (im guessing thats unlikely as it would struggle to interpret whats new in the location etc). If this is the case that it uses the same snapshot it will most likely fail at the same file?

    It would be awesome if when the job were to fail, it would mark the file with an archive bit or something to that affect. Then you could rerun it it would do a new snap and complete the job assessing archive bits to assess what needs to be retried.

    Kind Regards,

    Stuart

  • Usually a corrupt file is not really corrupt, but the file is opened. You should enable AOF in your job to handle opened files
  • Thats exactly what we are finding, the files are not really corrupt. 

     

    Thanks for the advice. 

     

    Just out of personal experience, what is maximum size jobs do you deal with? We find jobs hitting 2tb mark fail frequently, but we do get away with taking on 11-15 tb sometimes. 

     

    Regards,

    Stuart

  • My experience would have no bearing on your jobs. Each environment is different.