From 76ce3a707f8d77a36b1003d6fee9823a14376f3b Mon Sep 17 00:00:00 2001 From: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> Date: Tue, 25 Feb 2025 16:08:24 +0200 Subject: [PATCH 01/14] Update snapshots.mdx --- fusion_docs/guide/snapshots.mdx | 35 +++++++++++++++++++++++++++------ 1 file changed, 29 insertions(+), 6 deletions(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index 31c0ac278..1cb8b63fd 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -18,7 +18,7 @@ More specifically, the first use case for this feature is for Seqera Platform us Fusion Snapshots v1.0.0 requires the following [Seqera compute environment](https://docs.seqera.io/platform/latest/compute-envs/aws-batch) configuration: - **Provider**: AWS Batch -- **Pipeline work directory**: An S3 bucket in the same region as the compute environment +- **Pipeline work directory**: An S3 bucket located in the same region as your AWS Batch compute resources - **Enable Wave containers** - **Enable Fusion v2** - **Enable fast instance storage** @@ -43,18 +43,41 @@ fusion.containerConfigUrl = '' `maxSpotAttempts` must be a value higher than `0`. -### Recommended instance sizes +### EC2 instance selection guidelines -Fusion Snapshots require EC2 Spot instances with enough memory and network bandwidth to dump the cache of task intermediate files to S3 storage before AWS terminates an instance. When AWS issues a Spot instance reclamation notice, Fusion has two minutes to complete this transfer. +- Choose EC2 Spot instances with sufficient memory and network bandwidth to dump the cache of task intermediate files to S3 storage before AWS terminates an instance +- Select instances with guaranteed network bandwidth (not instances with bandwidth "up to" a maximum value) +- Maintain a 5:1 ratio between memory (GiB) and network bandwidth (Gbps) +- Recommended instance families: `c6id`, `r6id`, or `m6id` series instances work optimally with Fusion fast instance storage -It is recommended to select instances with guaranteed network bandwidth (as opposed to bandwidth _up to_ a maximum value) and maintain a ratio of 5:1 between memory and network bandwidth. +:::info "Example" +A c6id.8xlarge instance provides 64 GiB memory and 12.5 Gbps guaranteed network bandwidth. This configuration can transfer the entire memory contents to S3 in approximately 70 seconds, well within the 2-minute reclamation window. +::: + +:::note +Instances with lower network-to-memory ratios may not complete transfers before termination, potentially resulting in task failures. +::: -For example, taking into account the bandwidth and compute necessary to create a snapshot, a `c6i.8xlarge` instance with 64 GIB memory and a guaranteed network bandwidth of 12.5 Gbps can take approximately 70 seconds to dump the entire instance to S3 storage before instance reclamation occurs. +#### Recommended instance types -### Amazon Linux 2023 ECS-optimized AMI +| Instance Type | Memory (GiB) | Network Bandwidth (Gbps) | Memory:Bandwidth Ratio | Est. Snapshot Time | +|----------------|--------------|--------------------------|------------------------|-------------------| +| c6id.4xlarge | 32 | 12.5 | 2.56:1 | ~45 seconds | +| c6id.8xlarge | 64 | 12.5 | 5.12:1 | ~70 seconds | +| r6id.2xlarge | 64 | 12.5 | 5.12:1 | ~70 seconds | +| m6id.4xlarge | 64 | 12.5 | 5.12:1 | ~70 seconds | +| c6id.12xlarge | 96 | 18.75 | 5.12:1 | ~70 seconds | +| r6id.4xlarge | 128 | 12.5 | 10.24:1 | ~105 seconds | +| m6id.8xlarge | 128 | 25 | 5.12:1 | ~70 seconds | + +### Platform Enterprise: Select an Amazon Linux 2023 ECS-optimized AMI To obtain sufficient performance, Fusion Snapshots require instances with Amazon Linux 2023 (which ships with Linux Kernel 6.1), with an ECS Container-optimized AMI. +:::note +Selecting an Amazon Linux 2023 ECS-Optimized AMI is only required for Seqera Enterprise instances. Seqera Cloud AWS Batch compute environments select this AMI by default. +::: + To find the recommended AL2023 ECS-optimized AMI for your region, run the following (replace `eu-central-1` with your AWS region): ```bash From f7fefbb618b2aa26fde846a35d6b4e69f9dce941 Mon Sep 17 00:00:00 2001 From: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> Date: Tue, 25 Feb 2025 16:11:30 +0200 Subject: [PATCH 02/14] Update fusion_docs/guide/snapshots.mdx Signed-off-by: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> --- fusion_docs/guide/snapshots.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index 1cb8b63fd..c93f29263 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -70,7 +70,7 @@ Instances with lower network-to-memory ratios may not complete transfers before | r6id.4xlarge | 128 | 12.5 | 10.24:1 | ~105 seconds | | m6id.8xlarge | 128 | 25 | 5.12:1 | ~70 seconds | -### Platform Enterprise: Select an Amazon Linux 2023 ECS-optimized AMI +### (Seqera Enterprise only) Select an Amazon Linux 2023 ECS-optimized AMI To obtain sufficient performance, Fusion Snapshots require instances with Amazon Linux 2023 (which ships with Linux Kernel 6.1), with an ECS Container-optimized AMI. From 9839a0bcc57689beb3e3a8f71275f8358308faae Mon Sep 17 00:00:00 2001 From: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> Date: Tue, 25 Feb 2025 16:12:20 +0200 Subject: [PATCH 03/14] Update fusion_docs/guide/snapshots.mdx Signed-off-by: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> --- fusion_docs/guide/snapshots.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index c93f29263..bdb1ad1fd 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -75,7 +75,7 @@ Instances with lower network-to-memory ratios may not complete transfers before To obtain sufficient performance, Fusion Snapshots require instances with Amazon Linux 2023 (which ships with Linux Kernel 6.1), with an ECS Container-optimized AMI. :::note -Selecting an Amazon Linux 2023 ECS-Optimized AMI is only required for Seqera Enterprise instances. Seqera Cloud AWS Batch compute environments select this AMI by default. +Selecting an Amazon Linux 2023 ECS-Optimized AMI is only required for compute environments in Seqera Enterprise deployments. Seqera Cloud AWS Batch compute environments use this AMI by default. ::: To find the recommended AL2023 ECS-optimized AMI for your region, run the following (replace `eu-central-1` with your AWS region): From b5c223307ac350d4a80800d498e1404f606205fd Mon Sep 17 00:00:00 2001 From: Justine Geffen Date: Tue, 25 Feb 2025 16:23:34 +0200 Subject: [PATCH 04/14] Update snapshots.mdx Added punctuation to a list. Signed-off-by: Justine Geffen --- fusion_docs/guide/snapshots.mdx | 10 +++++----- 1 file changed, 5 insertions(+), 5 deletions(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index bdb1ad1fd..b50779985 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -45,10 +45,10 @@ fusion.containerConfigUrl = '' ### EC2 instance selection guidelines -- Choose EC2 Spot instances with sufficient memory and network bandwidth to dump the cache of task intermediate files to S3 storage before AWS terminates an instance -- Select instances with guaranteed network bandwidth (not instances with bandwidth "up to" a maximum value) -- Maintain a 5:1 ratio between memory (GiB) and network bandwidth (Gbps) -- Recommended instance families: `c6id`, `r6id`, or `m6id` series instances work optimally with Fusion fast instance storage +- Choose EC2 Spot instances with sufficient memory and network bandwidth to dump the cache of task intermediate files to S3 storage before AWS terminates an instance. +- Select instances with guaranteed network bandwidth (not instances with bandwidth "up to" a maximum value). +- Maintain a 5:1 ratio between memory (GiB) and network bandwidth (Gbps). +- Recommended instance families: `c6id`, `r6id`, or `m6id` series instances work optimally with Fusion fast instance storage. :::info "Example" A c6id.8xlarge instance provides 64 GiB memory and 12.5 Gbps guaranteed network bandwidth. This configuration can transfer the entire memory contents to S3 in approximately 70 seconds, well within the 2-minute reclamation window. @@ -100,4 +100,4 @@ The result for the `eu-central-1` region is similar to the following: } ``` -Note the `image_id` in your result (in this example, `ami-0281c9a5cd9de63bd`). Specify this ID in the **AMI ID** field under **Advanced options** when you create your Seqera compute environment. \ No newline at end of file +Note the `image_id` in your result (in this example, `ami-0281c9a5cd9de63bd`). Specify this ID in the **AMI ID** field under **Advanced options** when you create your Seqera compute environment. From b0913b62057c630706f2ab4c443f55f9ea666ea3 Mon Sep 17 00:00:00 2001 From: Justine Geffen Date: Tue, 25 Feb 2025 16:24:41 +0200 Subject: [PATCH 05/14] Update snapshots.mdx Table capitalization Signed-off-by: Justine Geffen --- fusion_docs/guide/snapshots.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index b50779985..bcba6f272 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -60,7 +60,7 @@ Instances with lower network-to-memory ratios may not complete transfers before #### Recommended instance types -| Instance Type | Memory (GiB) | Network Bandwidth (Gbps) | Memory:Bandwidth Ratio | Est. Snapshot Time | +| Instance type | Memory (GiB) | Network bandwidth (Gbps) | Memory:Bandwidth ratio | Est. snapshot time| |----------------|--------------|--------------------------|------------------------|-------------------| | c6id.4xlarge | 32 | 12.5 | 2.56:1 | ~45 seconds | | c6id.8xlarge | 64 | 12.5 | 5.12:1 | ~70 seconds | From e9b9cc86479f2803dd9d7a4c0d7863383f75fd0d Mon Sep 17 00:00:00 2001 From: Justine Geffen Date: Tue, 25 Feb 2025 16:26:01 +0200 Subject: [PATCH 06/14] Update fusion_docs/guide/snapshots.mdx Signed-off-by: Justine Geffen --- fusion_docs/guide/snapshots.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index bcba6f272..1815cdaeb 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -75,7 +75,7 @@ Instances with lower network-to-memory ratios may not complete transfers before To obtain sufficient performance, Fusion Snapshots require instances with Amazon Linux 2023 (which ships with Linux Kernel 6.1), with an ECS Container-optimized AMI. :::note -Selecting an Amazon Linux 2023 ECS-Optimized AMI is only required for compute environments in Seqera Enterprise deployments. Seqera Cloud AWS Batch compute environments use this AMI by default. +Selecting an Amazon Linux 2023 ECS-optimized AMI is only required for compute environments in Seqera Enterprise deployments. Seqera Cloud AWS Batch compute environments use this AMI by default. ::: To find the recommended AL2023 ECS-optimized AMI for your region, run the following (replace `eu-central-1` with your AWS region): From 5c083234daf739a9977be176a92054275162258e Mon Sep 17 00:00:00 2001 From: Justine Geffen Date: Tue, 25 Feb 2025 16:26:21 +0200 Subject: [PATCH 07/14] Update fusion_docs/guide/snapshots.mdx Signed-off-by: Justine Geffen --- fusion_docs/guide/snapshots.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index 1815cdaeb..7d82f41c5 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -60,7 +60,7 @@ Instances with lower network-to-memory ratios may not complete transfers before #### Recommended instance types -| Instance type | Memory (GiB) | Network bandwidth (Gbps) | Memory:Bandwidth ratio | Est. snapshot time| +| Instance type | Memory (GiB) | Network bandwidth (Gbps) | Memory:Bandwidth ratio | Est. Snapshot time| |----------------|--------------|--------------------------|------------------------|-------------------| | c6id.4xlarge | 32 | 12.5 | 2.56:1 | ~45 seconds | | c6id.8xlarge | 64 | 12.5 | 5.12:1 | ~70 seconds | From 3027cbcf289f13cfa5186dcf0e0656b81781ce27 Mon Sep 17 00:00:00 2001 From: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> Date: Tue, 25 Feb 2025 17:06:20 +0200 Subject: [PATCH 08/14] Update sidebar.json --- fusion_docs/sidebar.json | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fusion_docs/sidebar.json b/fusion_docs/sidebar.json index 67827cbdf..fcd510a50 100644 --- a/fusion_docs/sidebar.json +++ b/fusion_docs/sidebar.json @@ -23,7 +23,8 @@ "guide/local/minio", "guide/local/oracle-object" ] - } + }, + "guide/snapshots" ] }, "reference", From c6454a889844b1562d8741e3ea99cb568213a409 Mon Sep 17 00:00:00 2001 From: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> Date: Tue, 25 Feb 2025 17:09:23 +0200 Subject: [PATCH 09/14] Update sidebar.json --- fusion_docs/sidebar.json | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/fusion_docs/sidebar.json b/fusion_docs/sidebar.json index fcd510a50..67827cbdf 100644 --- a/fusion_docs/sidebar.json +++ b/fusion_docs/sidebar.json @@ -23,8 +23,7 @@ "guide/local/minio", "guide/local/oracle-object" ] - }, - "guide/snapshots" + } ] }, "reference", From efd357a97516b5ec291670f430a3f3781eb24cd3 Mon Sep 17 00:00:00 2001 From: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> Date: Tue, 25 Feb 2025 17:10:43 +0200 Subject: [PATCH 10/14] Update snapshots.mdx --- fusion_docs/guide/snapshots.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index 7d82f41c5..1e861414f 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -50,7 +50,7 @@ fusion.containerConfigUrl = '' - Maintain a 5:1 ratio between memory (GiB) and network bandwidth (Gbps). - Recommended instance families: `c6id`, `r6id`, or `m6id` series instances work optimally with Fusion fast instance storage. -:::info "Example" +:::info Example A c6id.8xlarge instance provides 64 GiB memory and 12.5 Gbps guaranteed network bandwidth. This configuration can transfer the entire memory contents to S3 in approximately 70 seconds, well within the 2-minute reclamation window. ::: From fb033ce2493cf06e48ca54b875d10bd6e9590411 Mon Sep 17 00:00:00 2001 From: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> Date: Tue, 25 Feb 2025 17:14:28 +0200 Subject: [PATCH 11/14] Update snapshots.mdx --- fusion_docs/guide/snapshots.mdx | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index 1e861414f..42799a98f 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -52,10 +52,8 @@ fusion.containerConfigUrl = '' :::info Example A c6id.8xlarge instance provides 64 GiB memory and 12.5 Gbps guaranteed network bandwidth. This configuration can transfer the entire memory contents to S3 in approximately 70 seconds, well within the 2-minute reclamation window. -::: -:::note -Instances with lower network-to-memory ratios may not complete transfers before termination, potentially resulting in task failures. +Instances with network-to-memory ratios lower than 5:1 may not complete transfers before termination, potentially resulting in task failures. ::: #### Recommended instance types From c10a0d6335882c291fd01b2ff634e1486a636336 Mon Sep 17 00:00:00 2001 From: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> Date: Tue, 25 Feb 2025 17:16:33 +0200 Subject: [PATCH 12/14] Update snapshots.mdx --- fusion_docs/guide/snapshots.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index 42799a98f..03c5a639f 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -53,7 +53,7 @@ fusion.containerConfigUrl = '' :::info Example A c6id.8xlarge instance provides 64 GiB memory and 12.5 Gbps guaranteed network bandwidth. This configuration can transfer the entire memory contents to S3 in approximately 70 seconds, well within the 2-minute reclamation window. -Instances with network-to-memory ratios lower than 5:1 may not complete transfers before termination, potentially resulting in task failures. +Instances with memory-to-network ratios over 5:1 may not complete transfers before termination, potentially resulting in task failures. ::: #### Recommended instance types From ce889dba1dd75961b5fd41fafbd1d62ef7a47060 Mon Sep 17 00:00:00 2001 From: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> Date: Tue, 25 Feb 2025 17:18:11 +0200 Subject: [PATCH 13/14] Improve Enterprise AL2023 note --- fusion_docs/guide/snapshots.mdx | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index 03c5a639f..b5f0d913a 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -73,7 +73,7 @@ Instances with memory-to-network ratios over 5:1 may not complete transfers befo To obtain sufficient performance, Fusion Snapshots require instances with Amazon Linux 2023 (which ships with Linux Kernel 6.1), with an ECS Container-optimized AMI. :::note -Selecting an Amazon Linux 2023 ECS-optimized AMI is only required for compute environments in Seqera Enterprise deployments. Seqera Cloud AWS Batch compute environments use this AMI by default. +Selecting a custom Amazon Linux 2023 ECS-optimized AMI is only required for compute environments in Seqera Enterprise deployments. Seqera Cloud AWS Batch compute environments use Amazon Linux 2023 AMIs by default. ::: To find the recommended AL2023 ECS-optimized AMI for your region, run the following (replace `eu-central-1` with your AWS region): From a7ba182787e187f48f6536d0aec6463bb77c6d52 Mon Sep 17 00:00:00 2001 From: Llewellyn vd Berg <113503285+llewellyn-sl@users.noreply.github.com> Date: Tue, 25 Feb 2025 17:24:50 +0200 Subject: [PATCH 14/14] Minor capitalization and note fix --- fusion_docs/guide/snapshots.mdx | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/fusion_docs/guide/snapshots.mdx b/fusion_docs/guide/snapshots.mdx index b5f0d913a..88d9294eb 100644 --- a/fusion_docs/guide/snapshots.mdx +++ b/fusion_docs/guide/snapshots.mdx @@ -53,12 +53,12 @@ fusion.containerConfigUrl = '' :::info Example A c6id.8xlarge instance provides 64 GiB memory and 12.5 Gbps guaranteed network bandwidth. This configuration can transfer the entire memory contents to S3 in approximately 70 seconds, well within the 2-minute reclamation window. -Instances with memory-to-network ratios over 5:1 may not complete transfers before termination, potentially resulting in task failures. +Instances with memory:bandwitdth ratios over 5:1 may not complete transfers before termination, potentially resulting in task failures. ::: #### Recommended instance types -| Instance type | Memory (GiB) | Network bandwidth (Gbps) | Memory:Bandwidth ratio | Est. Snapshot time| +| Instance type | Memory (GiB) | Network bandwidth (Gbps) | Memory:bandwidth ratio | Est. Snapshot time| |----------------|--------------|--------------------------|------------------------|-------------------| | c6id.4xlarge | 32 | 12.5 | 2.56:1 | ~45 seconds | | c6id.8xlarge | 64 | 12.5 | 5.12:1 | ~70 seconds |