Netbackup test
On 12/08/10 11:43 PM, Buchan Milne wrote:
The modifications is really a way of displaying the result and track "none 0" DONE jobs
It will keep track of jobs as they running it is organized as:
DUE jobs, upcoming jobs (not queued)
QUEUED Jobs,
ACTIVE Jobs,
DONE Jobs,
It will sort of track a job as it appears on DUE, moves down to QUEUED, when run it is on ACTIVE and finally ends up in DONE.
Any DONE jobs that are not 0 will go red and the page will go red, there is a function to "Ack" this by adding the NBU Job nr to a file.
I got a functionality to try to see if the backup is running or not at certain times, it works 98% of the times...
It will ONLY handle the bpdbjobs output from the NBU server it runs on
In our environment it takes some time to run bpdbjobs (30-60sec), and this is quite a small env.
I don't know what time it will take to run in a large env. This could be a problem if you run the test every 5min.
Maybe it is just suitable for my env.
My intention is to put it on Xymonton, if it's useful.
This script does what the originally nbu.ksh did, mangles the output of bpdbjobsOn Wednesday, 4 August 2010 07:30:39 Roland Soderstrom wrote:Hi, I got this heavily modified nbu.ksh script originally from deadcat NBU.3.4.1.02. It started as a very specialized test for our particular environment.We have a script that we have been running for some time, but it more or less just mangles the output of 'bperror -U -backstat -hoursago 72' or similar, and tries to extract the hostname, and send a report per client, with very crude "color" assignment, and the full output as a report for the master server. This method often gets things wrong, e.g. if a full backup and all but one incrememental backup succeeded the status color will be the same as if all backups failed. So, parsing the output in perl and assessing by policy and ensuring that at least one schedule for each policy succeeded may work better. And I guess -U shouldn't be used, otherwise client names and policy names get cut off.
The modifications is really a way of displaying the result and track "none 0" DONE jobs
It will keep track of jobs as they running it is organized as:
DUE jobs, upcoming jobs (not queued)
QUEUED Jobs,
ACTIVE Jobs,
DONE Jobs,
It will sort of track a job as it appears on DUE, moves down to QUEUED, when run it is on ACTIVE and finally ends up in DONE.
Any DONE jobs that are not 0 will go red and the page will go red, there is a function to "Ack" this by adding the NBU Job nr to a file.
I got a functionality to try to see if the backup is running or not at certain times, it works 98% of the times...
I just cleaned up certain 'just for us' settings and made them more configurable.I have modified it today to get a more common approach.Can you discuss the approach?
I guess it will handle any large environment but as it relies on bpdbjobs output the webpage could be quite large, as in many rows of bpdbjobs.It might not suit the enterprise NetBackup solutions as they might be to large. Smaller and midsized solutions might find it useful.What do you consider large? We're currently
It will ONLY handle the bpdbjobs output from the NBU server it runs on
In our environment it takes some time to run bpdbjobs (30-60sec), and this is quite a small env.
I don't know what time it will take to run in a large env. This could be a problem if you run the test every 5min.
The main reason to ask if anyone could test it is just to see if it is useful for anyone else.I use it for our Netbackup 6.5.4 on solaris with Xymon 4.3.0.0_beta2 Anyone interested to have a go? mail me at user-0cec9512a49f@xymon.invalid and I send a copy.
Maybe it is just suitable for my env.
My intention is to put it on Xymonton, if it's useful.
At present I have two similar environments. Regards, Buchan