# Linkerd2-cni: After upgrading to v30.8.5 cni pod crashes and can not be restarted

**URL:** <https://linkerd.buoyant.io/t/linkerd2-cni-after-upgrading-to-v30-8-5-cni-pod-crashes-and-can-not-be-restarted/407>\
**Category:** Linkerd General Discussion\
**Created:** [March 6, 2024, 7:45am UTC](https://linkerd.buoyant.io/t/linkerd2-cni-after-upgrading-to-v30-8-5-cni-pod-crashes-and-can-not-be-restarted/407 "2024-03-06T07:45:09Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![Hanife](https://avatars.discourse-cdn.com/v4/letter/h/a4c791/32.png) [@Hanife](https://linkerd.buoyant.io/u/Hanife)\
**Post date:** [March 6, 2024, 7:45am UTC](https://linkerd.buoyant.io/t/linkerd2-cni-after-upgrading-to-v30-8-5-cni-pod-crashes-and-can-not-be-restarted/407/1 "2024-03-06T07:45:09Z")

</div>

I have upgraded linkerd-crds and linkerd-control-plane to v2.13.7 . To get the compatible version of linkerd2-cni with other components , i have also upgraded it to v30.8.5 . But **linkerd-cni-29zxr** pod gives error and crashes. From pod logs :  
`/host/opt/cni/bin is non-writeable, failure

Stream closed EOF for linkerd-cni/linkerd-cni-29zxr (install-cni)`

 ![Screenshot 2024-03-06 at 08.43.13](https://canada1.discourse-cdn.com/flex029/uploads/buoyant1/original/1X/bd6f16f9efb0563d503549c976219833918a5494.jpeg)

---

<div class="post-metadata">

**Author:** ![alpeb](https://yyz1.discourse-cdn.com/flex029/user_avatar/linkerd.buoyant.io/alpeb/32/61_2.png) [@alpeb](https://linkerd.buoyant.io/u/alpeb)\
**Post date:** [March 6, 2024, 11:02am UTC](https://linkerd.buoyant.io/t/linkerd2-cni-after-upgrading-to-v30-8-5-cni-pod-crashes-and-can-not-be-restarted/407/2 "2024-03-06T11:02:50Z")

</div>

Hi, can you please provide more details like your k8s version and whether you’re using a cloud provider?

---

<div class="post-metadata">

**Author:** ![Hanife](https://avatars.discourse-cdn.com/v4/letter/h/a4c791/32.png) [@Hanife](https://linkerd.buoyant.io/u/Hanife)\
**Post date:** [March 6, 2024, 11:15am UTC](https://linkerd.buoyant.io/t/linkerd2-cni-after-upgrading-to-v30-8-5-cni-pod-crashes-and-can-not-be-restarted/407/3 "2024-03-06T11:15:02Z")

</div>

Hi, K8s v1.23.4, OTC cloud.

---

<div class="post-metadata">

**Author:** ![alpeb](https://yyz1.discourse-cdn.com/flex029/user_avatar/linkerd.buoyant.io/alpeb/32/61_2.png) [@alpeb](https://linkerd.buoyant.io/u/alpeb)\
**Post date:** [March 6, 2024, 2:19pm UTC](https://linkerd.buoyant.io/t/linkerd2-cni-after-upgrading-to-v30-8-5-cni-pod-crashes-and-can-not-be-restarted/407/4 "2024-03-06T14:19:01Z")

</div>

As shown in [these examples](https://github.com/linkerd/linkerd2/issues/2174), different providers place their CNI binaries in different places, that you can specify via the `destCNIBinDir` settings for the linkerd2-cni chart. We have no experience with OTC, so maybe you can help us find where those dirs are, searching for “cni” through the node’s file system, or asking your provider how they configured k8s , in particular the kubelet flag `--cni-bin-dir` value that was used.

---

<div class="post-metadata">

**Author:** ![Hanife](https://avatars.discourse-cdn.com/v4/letter/h/a4c791/32.png) [@Hanife](https://linkerd.buoyant.io/u/Hanife)\
**Post date:** [March 6, 2024, 3:58pm UTC](https://linkerd.buoyant.io/t/linkerd2-cni-after-upgrading-to-v30-8-5-cni-pod-crashes-and-can-not-be-restarted/407/5 "2024-03-06T15:58:21Z")

</div>

Hello again, the CNI binary in OTC Nodes placed also in /opt/cni/bin , it was not the reason of the issue. After configuring the DaemonSet with SecComp `Unconfined` , the update worked. Thank you for your Reply 👍