# Question about failure in coodinator

**URL:** <https://forums.digi.com/t/question-about-failure-in-coodinator/1850>\
**Category:** XBee - 802.15.4\
**Created:** [June 1, 2012, 12:47pm UTC](https://forums.digi.com/t/question-about-failure-in-coodinator/1850 "2012-06-01T12:47:14Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![titangui](https://avatars.discourse-cdn.com/v4/letter/t/7993a0/32.png) [@titangui](https://forums.digi.com/u/titangui)\
**Post date:** [June 1, 2012, 12:47pm UTC](https://forums.digi.com/t/question-about-failure-in-coodinator/1850/1 "2012-06-01T12:47:14Z")

</div>

Hi all,

I have one question about possible failures in XBee networks. I have one XBee network with coordinator and end devices working properly in API = 2. My coordinator send frames to end devices requesting information about sensors connected at the end devices. My question is if happen failure in coordinator and the coordinator need be replaced, is possible replace the broken coordinator for another coordinator keeping the network PAN ID and CHANNEL without need change any configuration or reset the end devices? Is possible put two coordinator in one network and if one failure the other takes the front? Anyone have suggestions of how make my network more tolerant a failure in coordinator?

Thanks.

---

<div class="post-metadata">

**Author:** ![lynnl](https://avatars.discourse-cdn.com/v4/letter/l/22d042/32.png) [@lynnl](https://forums.digi.com/u/lynnl)\
**Post date:** [June 5, 2012, 6:38pm UTC](https://forums.digi.com/t/question-about-failure-in-coodinator/1850/2 "2012-06-05T18:38:25Z")

</div>

This is the 802.15.4 forum, so I assume you have the S1, not S2/ZigBee.

You can have 2 ‘coordinators’ on the same channel even with the same PAN ID, but they will create distinct, competing networks and end-devcies will join one or the other somewhat randomly. If either coordinator fails, eventually the end-devices would migrate to the survining one. Some people do this and design their system not to care WHICH of the 2 coordinators the nodes join.

However, this would only work if the end-devices ‘broadcast’ their data back to the ‘master’, since if you unicast, then the DH/DL within the end-device would need to change if the destination changes.