cancel
Showing results for 
Search instead for 
Did you mean: 

Check the waiting time for backup

smith_jones
Level 6

Hi,

Can you please tell me what should I do if I want to increase the waiting time for backups?

Suppose there is one drive and one backup running. Second backup job is in queue since there is only one drive in this case.

After some specified time the second backup that was in queue will fail with error 196, since till that threshold time drive was not free.

So if I want to modify\increase that timing so that backup not fails (atleast one or two hours more), where to proceed?

Thanks.

1 ACCEPTED SOLUTION

Accepted Solutions

Genericus
Moderator
Moderator
   VIP   

If the window is 10 - 14, that is 6 hours. if it runs at 10, the frequency will tell it to start after 8 hours, which would be 1800. Since that is past the end of the start window, it will not run.

If it runs at 11, 11+8 = 19, will not run. since it is 8 hours not 1 day, next day will start at 10, not 11...

If the frequency is set at 2 hours, it will run at 10, 12, 14. ( I do that to run repeating backups for my oracle archive logs )

 

check the length of the window, make the frequency longer 

NetBackup 9.1.0.1 on Solaris 11, writing to Data Domain 9800 7.7.4.0
duplicating via SLP to LTO5 & LTO8 in SL8500 via ACSLS

View solution in original post

12 REPLIES 12

Genericus
Moderator
Moderator
   VIP   

You have three options, depending on which way you are scheduling it.

If frequency based:

A. expand the Start Window range so it can run longer - NOTE - you need to check the Frequency closely, if your DURATION goes past the minimal frequency it can start ANOTHER backup!

If calendar based:

A. set the schedule as calendar based, and select "Retries allowed after runday" on the Attributes tab (if it goes past midnight)

B. expand the Start Window range so it can run longer - NOTE - you need to check the days allowed to run, calendar scheduled that go past midnight can start ANOTHER backup for the next day...

 

Forget about it - just right click the 196 and restart it from the activity monitor...

 

NetBackup 9.1.0.1 on Solaris 11, writing to Data Domain 9800 7.7.4.0
duplicating via SLP to LTO5 & LTO8 in SL8500 via ACSLS

Marianne
Moderator
Moderator
Partner    VIP    Accredited Certified

Other options:

Implement Multiplexing for Storage Unit and Schedules

Select Allow Multiple Data Streams in Policy Attributes

Increase Max Jobs per Client in Master Server Global Attributes

Genericus
Moderator
Moderator
   VIP   

Good outside the box answer Marianne, you are absolutely correct - you can write both at once!

 

An additional answer might be to add disk storage in between - if you backup to disk, it is often faster, and you can write multiple images at once without multiplexing.

Usiung SLP, you can then duplicate the images to tape, without having to worry about backup windows, or impacting production.

 

It really will depend on the environment - if the source disk is slow, writing two at once generally means now you have two running half as fast...

If the tape drive is the weak link, multiplexing or writing two at once to disk may be the solution.

 

 

NetBackup 9.1.0.1 on Solaris 11, writing to Data Domain 9800 7.7.4.0
duplicating via SLP to LTO5 & LTO8 in SL8500 via ACSLS

smith_jones
Level 6

Hi,

@Genericus:

For your answer -

If frequency based:

A. expand the Start Window range so it can run longer - NOTE - you need to check the Frequency closely, if your DURATION goes past the minimal frequency it can start ANOTHER backup!

 As I know this backup window ( including start time and end time) is for starting the backup, right? Say backup is scheduled for 1 hour from 10:00 AM to 11:00 AM ( this is the backup window), then this backup will be in queue from 10:00AM to 11:00 AM if it not gets drive and will fail after that. But if it gets drive at time 10:55 AM it will start to run. Am I right or missing something in this concept?

So my concern is if backup window is available as per schedule and backup is in queue, then no issue, this backup job will remain in queue but if the backup window lapsed and then also I want the backup job to remain in queue then what should I do? 

I am just concerned for this because before yesterday everything was fine as per schedule but yesterday due to some urgent activities other backups like duplication were triggered manually which used drives. So I just want to know if in some scenario  a backup job is in queue more than its backup window and I am not there in office then how should I prevent failing this?

@Marianne:

I have those enabled but both the backups are totally different.

 

Thanks.

Genericus
Moderator
Moderator
   VIP   

If your backup start time window is 1000 - 1100, you can stretch that from 1000-1400, then set the frequency to 8 hours. Since the start window is not 8 hours long, it will not run twice, and will always start at the first opportunity (at 1000)

if calendar based, there is no problem to stretch the time, as long as you do not go past midnight.

 

I generally avoid using 1 day as frequency, use a number of hours that is longer than the start window instead.

 

NetBackup 9.1.0.1 on Solaris 11, writing to Data Domain 9800 7.7.4.0
duplicating via SLP to LTO5 & LTO8 in SL8500 via ACSLS

smith_jones
Level 6

Genericus,

Say I set the frequency to 8 hours. But this is good when the queued backup fails i.e. was not able to run in the specified backup window.

But if that queued backup gets the resource say for example drive (for which it was queued) in the specified backup window then it will be successfully completed. And since the frequency is set to 8 hours, it will again run after 8 hours and again after 8 hours. So consuming space. And this will be on daily basis.

Your answer is providing solution in one way but raising a major concern in second way. 

smith_jones
Level 6

I was running backup one time in one day but as per your solution I have to run it twice or thrice a day. 

Genericus
Moderator
Moderator
   VIP   

If the window is 10 - 14, that is 6 hours. if it runs at 10, the frequency will tell it to start after 8 hours, which would be 1800. Since that is past the end of the start window, it will not run.

If it runs at 11, 11+8 = 19, will not run. since it is 8 hours not 1 day, next day will start at 10, not 11...

If the frequency is set at 2 hours, it will run at 10, 12, 14. ( I do that to run repeating backups for my oracle archive logs )

 

check the length of the window, make the frequency longer 

NetBackup 9.1.0.1 on Solaris 11, writing to Data Domain 9800 7.7.4.0
duplicating via SLP to LTO5 & LTO8 in SL8500 via ACSLS

smith_jones
Level 6

Hi Genericus,

So frequency depends on the backup window. Really thanks for this clarity.

 

Genericus
Moderator
Moderator
   VIP   

Exactly, and thanks for marking a solution!

 

NetBackup 9.1.0.1 on Solaris 11, writing to Data Domain 9800 7.7.4.0
duplicating via SLP to LTO5 & LTO8 in SL8500 via ACSLS

fawad55
Level 3
Partner

please explain a little more i always confuse in backup window

smith_jones
Level 6

Hi fawad55,

You can think backup window as the time range that has been specified for the backup when creating the policy, i.e. the start time of the backup to end time of the backup ( the end time can also be treated as the last point of that backup to trigger if it was not triggered yet at the start time automatically).