# Is there truly no way to spread RF=3 data over 4 DC while torelarting a single datacenter failure?

**URL:** <https://forum.yugabyte.com/t/is-there-truly-no-way-to-spread-rf-3-data-over-4-dc-while-torelarting-a-single-datacenter-failure/4946>\
**Category:** General\
**Created:** [March 2, 2026, 6:13pm UTC](https://forum.yugabyte.com/t/is-there-truly-no-way-to-spread-rf-3-data-over-4-dc-while-torelarting-a-single-datacenter-failure/4946 "2026-03-02T18:13:05Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![hispebarzu](https://avatars.discourse-cdn.com/v4/letter/h/9de053/32.png) [@hispebarzu](https://forum.yugabyte.com/u/hispebarzu)\
**Post date:** [March 2, 2026, 6:13pm UTC](https://forum.yugabyte.com/t/is-there-truly-no-way-to-spread-rf-3-data-over-4-dc-while-torelarting-a-single-datacenter-failure/4946/1 "2026-03-02T18:13:05Z")

</div>

The only option I can think of is this, to force pin 1 tablet to dc1, and second to dc2, and leave third to place randomly it wants:

```auto
yb-admin --master-addresses=*** modify_placement_info datacenter1.*.*:1,datacenter2.*.*:1 3

```

Maybe there is some kind of anti-affinity for tablets placement?  
But this will lead to `Cluster Load is not Balanced` in my testing, which is true.

 ![rf_3_dc_4_uneven_distribution](https://canada1.discourse-cdn.com/flex027/uploads/yugabyte/original/2X/a/aff425c993439b9947d17b8fd80c35380dbdf1ca.png)

---

<div class="post-metadata">

**Author:** ![dorian\_yugabyte](https://yyz1.discourse-cdn.com/flex027/user_avatar/forum.yugabyte.com/dorian_yugabyte/32/206_2.png) [@dorian\_yugabyte](https://forum.yugabyte.com/u/dorian_yugabyte)\
**Post date:** [March 3, 2026, 8:50am UTC](https://forum.yugabyte.com/t/is-there-truly-no-way-to-spread-rf-3-data-over-4-dc-while-torelarting-a-single-datacenter-failure/4946/2 "2026-03-03T08:50:55Z")

</div>

Hi @hispebarzu

We have wildcard placement but we don’t have ant-affinity policy or a setting that guarantees that copies are placed as wide as possible. If needed, comment on [https://github.com/yugabyte/yugabyte-db/issues/19862](https://github.com/yugabyte/yugabyte-db/issues/19862)

Can you please describe your use case?

---

<div class="post-metadata">

**Author:** ![hispebarzu](https://avatars.discourse-cdn.com/v4/letter/h/9de053/32.png) [@hispebarzu](https://forum.yugabyte.com/u/hispebarzu)\
**Post date:** [March 6, 2026, 2:29pm UTC](https://forum.yugabyte.com/t/is-there-truly-no-way-to-spread-rf-3-data-over-4-dc-while-torelarting-a-single-datacenter-failure/4946/3 "2026-03-06T14:29:18Z")

</div>

This generally useful when we have limited vertical capacity, in the DC, but we source compute in multiple DC providers.