none
circular capture using nmcap RRS feed

Answers

  • Hi AJ,

    At this time we only support chained or circular mode but not a circular chain like you seem to be looking for.  I know others have written a script to run concurrently to delete older parts of a chain, but I think that's the only way you could do it currently.

    Thanks,


    Michael Hawker | Program Manager | Network Monitor

    Wednesday, August 22, 2012 6:24 PM
    Moderator

All replies

  • Hi AJ,

    At this time we only support chained or circular mode but not a circular chain like you seem to be looking for.  I know others have written a script to run concurrently to delete older parts of a chain, but I think that's the only way you could do it currently.

    Thanks,


    Michael Hawker | Program Manager | Network Monitor

    Wednesday, August 22, 2012 6:24 PM
    Moderator
  • Hi Michael,

    Thanks for the confirmation,  I hope this feature would be available in the future release.

    Regards,

    Aj

    Thursday, August 23, 2012 3:41 AM
  • AJ, it would be useful to understand the specific needs for your question.  As we work on future version, having the purpose behind the feature will help us design our product correctly.

    The main differences that I'm aware of between a circular capture and a chained buffer ring are that you can access the older files while the trace is running and that you have a larger total size.  If you had a circular buffer of a larger size, and could access a portion of the captured data while it was running, would that solve your problem?  If not can you explain how you would use this ring buffer today?

    Thanks,

    Paul

    Wednesday, August 29, 2012 2:53 PM
  • Hi Paul,

    Sorry for the late response.

    We were trying to use netmon to troubleshoot a problem where the backups fail intermittently, we could not use netmon with the current options as the disk gets filled up within couple of minutes and had to manually delete the oldest logs.

    The cause of the issue was environmental but we wanted to collect network traces (from the source and the destination nodes) to investigate the root cause. The option we were looking for was to set netmon to capture x number of logs of the size y MB(chained ring buffer) from command line which can be used to create a batch/script.

    This script can be used automate the network capture instead of having to manually delete older logs (as per the current design).

    For example lets assume the following environment:

    Server A, B: Windows Servers where the backups fail with network related errors

    Server C : backup destination for other servers

    To find out the root cause of the backup failure we have to collect simultaneous network captures in all the servers when the job is running and as the issue happens intermittently we need to automate the capture. With the current option of just circular capture we have to manually delete the oldest logs so that the disk doesnot get filled up.

    With Chained Ring Buffer the advantage is that at the start of the capture we can set the maximum number of logs, let’s say we set the maximum number of logs as 20 and the size of each log as 200 MB , when the 21st log is created it would overwrite the 1st logs so at any given point of time there would only be 20 logs and we know the disk would not fill up.

    We can now easily create a script to start the capture in all the three servers just before the backup job starts and when its fails run another script to stop the capture.

    Hope this helps in understanding what we were trying to achieve.

    Regards,

    Ajay

    Wednesday, September 26, 2012 1:44 AM