This question is really toward the development team for IPSLA unless someone else has run into this and can offer another idea?
We have been using IPSLA since its inception to monitor the voice quality of our network to several call centers. We have dedicated Cisco 7304 routers running in our two main data centers and have an IPSLA UDP Jitter operation set up to each of 60 or so destinations (i.e. hub and spoke). These operations have been added both using the automated and manual methods over a long period of time and everything had been working great up until we rebooted one of the 7304 routers.
After the reboot we got many errors on each path and narrowed the issue down to all of the IPSLA operations starting simultaneously. I wrote a simple script to restart each of the IPSLA operations and the problem cleared right up. It only requires a few millisenconds of separation to solve the issue.
Cisco is aware of this behaviour and have implemented a group start option to randomize the startup time of several operations over a preset period of time.
http://www.cisco.com/en/US/docs/ios/12_4t/12_4t2/ht_slars.html
At this point we are thinking our best bet is to convert all of the operations on these two routers to manual and set up the group start randomize option. Then go into Orion and delete all the existing stuff and re-add it using the manual discovery process. This way we can add unique operation ID's and keep track of them all.
Anyway it would be a nice feature in IPSLA if you could check an option to randomize the startup of all the operations once you exceed a preset number. Or just have a check box to tell it to create the operation in pending mode instead of start now. That way one would only need to add the group start command manually to each router effected.
Thanks.
