02-10-2015 11:57 PM
Hello All,
I have an issue with the Backup of Cluster Server. There are 22 file systems for each node . All the 22 streams are being triggered for one server. But only 16 streams are being triggered for the other server. This backup is successful. But the 6 streams which have not been started are being triggered at a later time( with a gap of 3 hours) and failing. I would like to know the reason why the 6 streams are not being triggered at the correct time.
Appreciate a quick response. Thanks in Advance
Solved! Go to Solution.
02-11-2015 03:12 AM
I honestly don't know how else to explain.... Have you read through the article that I have posted above?
NetBackup Frequency Based Scheduling
I am trying to tell you that jobs were probably queued some time ago - not last night. The 1st time that the jobs were queued was the day that the scheduling for those jobs were pushed out.
'Every one day' means that this stream will only be submitted 24 hours after previous successful start time.
The suggestion to change the frequency will solve your issue.
If you change the frequency to 13 or 14 hours, the jobs can be submitted 13 hours after the start time of previous successful backup. If the backup window is still closed, NBU will wait till the backup window opens and then submit the job.
Please post output of commands in my previous post.
This will tell us if the scheduling was pushed out recently (since 1 Feb).
02-11-2015 12:20 AM
If 'allow multiple data streams' is selected in the policy and Frequency based scheduling is used, then each stream (file system) will be evaluated based on the start time of the previous successful backup.
So, if 22 streams were started on day one at 18:00, but only resources to activate 16 streams and the rest queued, more resources available 3 hours later (21:00) to activate queued streams, then backup times will be adjusted on day 2 depending on Frequency setting.
Every 1 day means every 24 hours. So, the streams that went active at 21:00 on day one, will now have a start time 24 hours later on day 2 - 21:00 again.
We call this 'schedule creeping in Frequency scheduling and is overcome by making the frequency just slightly bigger than the backup window.
If backup window is 8 hours, change frequency to 9 hours.
Details in this article: NetBackup Frequency Based Scheduling
02-11-2015 12:39 AM
I understand that if the streams do not have the resource then they will start when they have the resources i.e the next day. But here the case is,they are being started on the same day.
16 streams of one server,22 streams of another server---- backup successful
After 3 hours
Only 6 streams of the first server are starting without any other streams -- backup failing
What can be the solution for the above case?
02-11-2015 12:49 AM
Are these 6 streams queueing in the activity monitor or not showing up whilst the other streams are running?
02-11-2015 12:55 AM
As per my previous post - it all depends on when the 'schedule kreeping' started. If some backups were queued some time ago and only activated couple of hours later, this is when then schedule for these streams got pushed out.
If you are using Frequency based scheduling, adjust the frequency as per my suggestion above.
Then check Activity Monitor when backup window opens to see the 'big picture'. How many jobs go active? How many jobs are queued? What is the reason for jobs being queued?
You never confirmed if you are indeed using Frequency scheduling?
And if so, what the frequency is set to?
To troubleshoot start time of frequency schedules, look at start times of previous successful jobs:
bpimagelist -client <client-name> -d 02/01/2015 -U
Then compare with schedules in policy config:
bppllist <policy-name> -U
02-11-2015 02:36 AM
These 6 streams are not even being shown in the queued list,they are not showing up while the other streams are running. Yes we are using the frequency based scheduling.The frequency is set to 1 day.
My backup window is 12 hours and 30 minutes and Frequency is set to 1 day(24 hours).I think this setting is fine??
02-11-2015 03:12 AM
I honestly don't know how else to explain.... Have you read through the article that I have posted above?
NetBackup Frequency Based Scheduling
I am trying to tell you that jobs were probably queued some time ago - not last night. The 1st time that the jobs were queued was the day that the scheduling for those jobs were pushed out.
'Every one day' means that this stream will only be submitted 24 hours after previous successful start time.
The suggestion to change the frequency will solve your issue.
If you change the frequency to 13 or 14 hours, the jobs can be submitted 13 hours after the start time of previous successful backup. If the backup window is still closed, NBU will wait till the backup window opens and then submit the job.
Please post output of commands in my previous post.
This will tell us if the scheduling was pushed out recently (since 1 Feb).
02-11-2015 03:57 AM
Please explain what you mean with (highlighted in bold below)
02-11-2015 05:15 AM
Hello All,
Thanks a lottt !!!!! for your quick responses and information.
From the post I can see that,if few backup jobs are not run,and if there is still time for a backup to be triggered(based on frequency check),then it will add a full backup to be run for those streams.
Those streams are being failed with an error code customized by us.
As I have limited access to the server couldn't paste the output of the commands.
Here policy has been configured in such a way that only weekly full backup will run on all the days. I guess this might be causing trouble. We have suggested the necessary changes to be done.
Will get back if the issue doesn't get solved even after the changes are made.
"I understand that if the streams do not have the resource then they will start when they have the resources i.e the next day. But here the case is,they are being started on the same day.
16 streams of one server,22 streams of another server---- backup successful
After 3 hours
Only 6 streams of the first server are starting without any other streams -- backup failing
What can be the solution for the above case?"
Failing with what error?