# Using mender with unreliable internet connection

**URL:** <https://hub.mender.io/t/using-mender-with-unreliable-internet-connection/5293>\
**Category:** General Discussions\
**Created:** [September 30, 2022, 9:30am UTC](https://hub.mender.io/t/using-mender-with-unreliable-internet-connection/5293 "2022-09-30T09:30:16Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![pulkomandy](https://yyz2.discourse-cdn.com/flex036/user_avatar/hub.mender.io/pulkomandy/32/1742_2.png) [@pulkomandy](https://hub.mender.io/u/pulkomandy)\
**Post date:** [September 30, 2022, 9:30am UTC](https://hub.mender.io/t/using-mender-with-unreliable-internet-connection/5293/1 "2022-09-30T09:30:16Z")

</div>

Hello,

In our setup, we use Mender to udpate devices which have an unreliable access to the internet.

We found several problems with this:

- If Mender-client cannot connect to the server after a reboot for some time, it will apparently automatically rollback, despite our artifact-commit script saying it’s OK to commit the changes. Is there a way to disable this?
- After this happens, if we want to attempt re-installing the same update, we have to download it to the device again, which may take several hours on our slow connection. Is there a way the mender-client could keep the latest uploaded artifact so it doesn’t need to re-download it in this case?

We will later be experimenting with delta updates to make the update files smaller, but I don’t know yet if this will be enough and it will certainly not fix the first problem.

Thanks for your help!

---

<div class="post-metadata">

**Author:** ![kacf](https://yyz2.discourse-cdn.com/flex036/user_avatar/hub.mender.io/kacf/32/146_2.png) [@kacf](https://hub.mender.io/u/kacf)\
**Post date:** [October 3, 2022, 7:44am UTC](https://hub.mender.io/t/using-mender-with-unreliable-internet-connection/5293/2 "2022-10-03T07:44:40Z")

</div>

> [@pulkomandy](#):
>
> If Mender-client cannot connect to the server after a reboot for some time, it will apparently automatically rollback, despite our artifact-commit script saying it’s OK to commit the changes. Is there a way to disable this?

Not currently. It would be quite dangerous to remove this check in general. There is no way for the client to know whether failure to connect is because of a legitimate network outage or a broken network driver in the new update, so it has to confirm.

I would recommend raising [`RetryPollIntervalSeconds`](https://docs.mender.io/client-installation/configuration-file/configuration-options#retrypollintervalseconds) instead, perhaps in combination with `RetryPollCount`.
