From 173b6a6a21512826b4f7d905c39f5535f9d681c3 Mon Sep 17 00:00:00 2001 From: Joongi Kim Date: Wed, 30 Nov 2022 17:37:24 +0000 Subject: [PATCH 1/3] fix: Misordering of cluster kernels in the host networking mode --- src/ai/backend/manager/registry.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/src/ai/backend/manager/registry.py b/src/ai/backend/manager/registry.py index 78bd6f38a24..2d03bf0c7b0 100644 --- a/src/ai/backend/manager/registry.py +++ b/src/ai/backend/manager/registry.py @@ -1304,7 +1304,7 @@ async def start_session( agent_addr = item.agent_alloc_ctx.agent_addr.replace( "tcp://", "" ).split(":", maxsplit=1)[0] - cluster_ssh_port_mapping[f"{cluster_role}{index+1}"] = ( + cluster_ssh_port_mapping[item.kernel.cluster_hostname] = ( agent_addr, port, ) From 04fc2cdcd2bb700e056fcea9e3040bde503533fe Mon Sep 17 00:00:00 2001 From: Joongi Kim Date: Wed, 30 Nov 2022 17:39:28 +0000 Subject: [PATCH 2/3] doc: Add news fragment --- changes/907.fix.md | 1 + 1 file changed, 1 insertion(+) create mode 100644 changes/907.fix.md diff --git a/changes/907.fix.md b/changes/907.fix.md new file mode 100644 index 00000000000..dfa572153c0 --- /dev/null +++ b/changes/907.fix.md @@ -0,0 +1 @@ +Fix occasionally random mismatch of cluster hostnames and actual container hostnames in cluster sessions under the host networking mode From 310eef4a0c8835abe1552e77e65df6e7f2e22ec3 Mon Sep 17 00:00:00 2001 From: Joongi Kim Date: Wed, 30 Nov 2022 17:41:53 +0000 Subject: [PATCH 3/3] doc: Update news fragment [skip ci] --- changes/907.fix.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/changes/907.fix.md b/changes/907.fix.md index dfa572153c0..7d43ce13e2f 100644 --- a/changes/907.fix.md +++ b/changes/907.fix.md @@ -1 +1 @@ -Fix occasionally random mismatch of cluster hostnames and actual container hostnames in cluster sessions under the host networking mode +Fix occasional random mismatch of cluster hostnames and actual container hostnames in cluster sessions under the host networking mode