2023-08-14 22:10:43 |
Mitchell Dzurick |
bug |
|
|
added bug |
2023-08-14 22:11:00 |
Mitchell Dzurick |
ec2-hibinit-agent (Ubuntu): assignee |
|
Mitchell Dzurick (mitchdz) |
|
2023-08-15 22:33:00 |
Mitchell Dzurick |
summary |
Add Swap Optimization from upstream |
Add recreating the swap file if the current size is sufficiently larger |
|
2023-08-28 13:50:57 |
Mitchell Dzurick |
summary |
Add recreating the swap file if the current size is sufficiently larger |
FFE+SRU: Add recreating the swap file if the current size is sufficiently larger |
|
2023-08-28 14:07:02 |
Mitchell Dzurick |
description |
Upstream added a swap optimization - https://github.com/aws/amazon-ec2-hibinit-agent/pull/31
Include this in the ubuntu package. |
Upstream Change - https://github.com/aws/amazon-ec2-hibinit-agent/pull/31
[FFE]
====
This change includes a small bugfix/feature addition to resize the swap partition if the user changes the AWS EC2 instance. On hibernation, the contents of RAM is put onto SWAP, so if a user creates an instance with a small amount of SWAP, and upgrades the instance, hibernation may fail.
[SRU]
====
[Impact]
If users create an instance with a small RAM/SWAP size and then upgrade the instance to an instance with more RAM, hibernation may not work as not enough SWAP will be provisioned.
This update dynamically allocates the SWAP so that if a user upgrades the instance then hibernation will still be available, and if the user downgrades an instance, SWAP will be resized to use less disk space.
[Test Case]
1. create & start a c5.xlarge instance.
2. Upgrade to the new ec2-hibinit-agent package.
3. Stop the instance.
4. Modified to a smaller instance (c5.large)
5. `journalctl -b | grep "Swap already exists!"` verify swap is recreated
Aug 04 18:19:03 ip-172-31-22-174.ec2.internal hibinit-agent[2457]: Swap already exists! (have 8011616256, need 4194304000), deleting existing swap file /swap since current swap is sufficiently large and wasting memory
7. Stop and restart the instance again
8. `journalctl -b | grep "There's sufficient swap available"` to verified no recreation. of swap occured.
Aug 04 18:23:17 ip-172-31-22-174.ec2.internal hibinit-agent[2452]: There's sufficient swap available (have 4194304000, need 4194304000)
9. Stop the instance
10. Modify to larger (c5.xlarge) instance.
11. Start the instance.
12. `journalctl -b | grep "Swap already exists!"` to verify swap is created due to larger swap required.
Aug 04 18:25:03 ip-172-31-22-174.ec2.internal hibinit-agent[2600]: Swap already exists! (have 4194304000, need 8011620352), deleting existing swap file /swap
[Where things could go wrong]
- If Dynamic swap allocation fails, then hibernation may not properly save the state of the machine.
- Perhaps the user relies on the smaller/bigger SWAP size for something else and dynamically resizing on instance change could cause issues? I'd say unlikely but could be a concern. |
|
2023-08-30 19:58:21 |
Launchpad Janitor |
merge proposal linked |
|
https://code.launchpad.net/~mitchdz/ubuntu/+source/ec2-hibinit-agent/+git/ec2-hibinit-agent/+merge/450277 |
|
2023-08-30 20:17:50 |
Mitchell Dzurick |
description |
Upstream Change - https://github.com/aws/amazon-ec2-hibinit-agent/pull/31
[FFE]
====
This change includes a small bugfix/feature addition to resize the swap partition if the user changes the AWS EC2 instance. On hibernation, the contents of RAM is put onto SWAP, so if a user creates an instance with a small amount of SWAP, and upgrades the instance, hibernation may fail.
[SRU]
====
[Impact]
If users create an instance with a small RAM/SWAP size and then upgrade the instance to an instance with more RAM, hibernation may not work as not enough SWAP will be provisioned.
This update dynamically allocates the SWAP so that if a user upgrades the instance then hibernation will still be available, and if the user downgrades an instance, SWAP will be resized to use less disk space.
[Test Case]
1. create & start a c5.xlarge instance.
2. Upgrade to the new ec2-hibinit-agent package.
3. Stop the instance.
4. Modified to a smaller instance (c5.large)
5. `journalctl -b | grep "Swap already exists!"` verify swap is recreated
Aug 04 18:19:03 ip-172-31-22-174.ec2.internal hibinit-agent[2457]: Swap already exists! (have 8011616256, need 4194304000), deleting existing swap file /swap since current swap is sufficiently large and wasting memory
7. Stop and restart the instance again
8. `journalctl -b | grep "There's sufficient swap available"` to verified no recreation. of swap occured.
Aug 04 18:23:17 ip-172-31-22-174.ec2.internal hibinit-agent[2452]: There's sufficient swap available (have 4194304000, need 4194304000)
9. Stop the instance
10. Modify to larger (c5.xlarge) instance.
11. Start the instance.
12. `journalctl -b | grep "Swap already exists!"` to verify swap is created due to larger swap required.
Aug 04 18:25:03 ip-172-31-22-174.ec2.internal hibinit-agent[2600]: Swap already exists! (have 4194304000, need 8011620352), deleting existing swap file /swap
[Where things could go wrong]
- If Dynamic swap allocation fails, then hibernation may not properly save the state of the machine.
- Perhaps the user relies on the smaller/bigger SWAP size for something else and dynamically resizing on instance change could cause issues? I'd say unlikely but could be a concern. |
Upstream Change - https://github.com/aws/amazon-ec2-hibinit-agent/pull/31
[FFE]
====
This change includes a small bugfix to resize the swap partition. On hibernation, the contents of RAM is put onto SWAP, so if a user upgrades the amount of RAM on an instance, SWAP needs to be recreated.
[SRU]
====
[Impact]
If an AWS EC2 user upgrades the amount of RAM on an instance, hibernation will fail.
This update dynamically allocates the SWAP to the necessary amount. This has the added benefit of downsizing SWAP on RAM reduction thus optimizing disk space in those scenarios.
[Test Case]
1. create & start a c5.xlarge instance.
2. Upgrade to the new ec2-hibinit-agent package.
3. Stop the instance.
4. Modified to a smaller instance (c5.large)
5. `journalctl -b | grep "Swap already exists!"` verify swap is recreated
Aug 04 18:19:03 ip-172-31-22-174.ec2.internal hibinit-agent[2457]: Swap already exists! (have 8011616256, need 4194304000), deleting existing swap file /swap since current swap is sufficiently large and wasting memory
7. Stop and restart the instance again
8. `journalctl -b | grep "There's sufficient swap available"` to verified no recreation. of swap occured.
Aug 04 18:23:17 ip-172-31-22-174.ec2.internal hibinit-agent[2452]: There's sufficient swap available (have 4194304000, need 4194304000)
9. Stop the instance
10. Modify to larger (c5.xlarge) instance.
11. Start the instance.
12. `journalctl -b | grep "Swap already exists!"` to verify swap is created due to larger swap required.
Aug 04 18:25:03 ip-172-31-22-174.ec2.internal hibinit-agent[2600]: Swap already exists! (have 4194304000, need 8011620352), deleting existing swap file /swap
[Where things could go wrong]
- If Dynamic swap allocation fails, then hibernation may not properly save the state of the machine.
- Perhaps the user relies on the smaller/bigger SWAP size for something else and dynamically resizing on instance change could cause issues? I'd say unlikely but could be a concern. |
|
2023-08-30 21:44:48 |
Mitchell Dzurick |
bug |
|
|
added subscriber Ubuntu Release Team |
2023-08-30 21:52:57 |
Mitchell Dzurick |
description |
Upstream Change - https://github.com/aws/amazon-ec2-hibinit-agent/pull/31
[FFE]
====
This change includes a small bugfix to resize the swap partition. On hibernation, the contents of RAM is put onto SWAP, so if a user upgrades the amount of RAM on an instance, SWAP needs to be recreated.
[SRU]
====
[Impact]
If an AWS EC2 user upgrades the amount of RAM on an instance, hibernation will fail.
This update dynamically allocates the SWAP to the necessary amount. This has the added benefit of downsizing SWAP on RAM reduction thus optimizing disk space in those scenarios.
[Test Case]
1. create & start a c5.xlarge instance.
2. Upgrade to the new ec2-hibinit-agent package.
3. Stop the instance.
4. Modified to a smaller instance (c5.large)
5. `journalctl -b | grep "Swap already exists!"` verify swap is recreated
Aug 04 18:19:03 ip-172-31-22-174.ec2.internal hibinit-agent[2457]: Swap already exists! (have 8011616256, need 4194304000), deleting existing swap file /swap since current swap is sufficiently large and wasting memory
7. Stop and restart the instance again
8. `journalctl -b | grep "There's sufficient swap available"` to verified no recreation. of swap occured.
Aug 04 18:23:17 ip-172-31-22-174.ec2.internal hibinit-agent[2452]: There's sufficient swap available (have 4194304000, need 4194304000)
9. Stop the instance
10. Modify to larger (c5.xlarge) instance.
11. Start the instance.
12. `journalctl -b | grep "Swap already exists!"` to verify swap is created due to larger swap required.
Aug 04 18:25:03 ip-172-31-22-174.ec2.internal hibinit-agent[2600]: Swap already exists! (have 4194304000, need 8011620352), deleting existing swap file /swap
[Where things could go wrong]
- If Dynamic swap allocation fails, then hibernation may not properly save the state of the machine.
- Perhaps the user relies on the smaller/bigger SWAP size for something else and dynamically resizing on instance change could cause issues? I'd say unlikely but could be a concern. |
Upstream Change - https://github.com/aws/amazon-ec2-hibinit-agent/commit/d2ce783d10d3a4437ebac468ed669666d8b6671b
[FFE]
====
This change includes a small bugfix to resize the swap partition. On hibernation, the contents of RAM is put onto SWAP, so if a user changes the amount of RAM on an instance, SWAP potentially needs to be recreated.
[SRU]
====
[Impact]
If an AWS EC2 user increases the amount of RAM on an instance, hibernation will fail.
This update dynamically allocates the SWAP to the necessary amount for the user. This has the added benefit of downsizing SWAP on RAM reduction thus optimizing disk space in those scenarios.
[Test Case]
1. create & start a c5.xlarge instance.
2. Upgrade to the new ec2-hibinit-agent package.
3. Stop the instance.
4. Modified to a smaller instance (c5.large)
5. `journalctl -b | grep "Swap already exists!"` verify swap is recreated
Aug 04 18:19:03 ip-172-31-22-174.ec2.internal hibinit-agent[2457]: Swap already exists! (have 8011616256, need 4194304000), deleting existing swap file /swap since current swap is sufficiently large and wasting memory
7. Stop and restart the instance again
8. `journalctl -b | grep "There's sufficient swap available"` to verified no recreation. of swap occured.
Aug 04 18:23:17 ip-172-31-22-174.ec2.internal hibinit-agent[2452]: There's sufficient swap available (have 4194304000, need 4194304000)
9. Stop the instance
10. Modify to larger (c5.xlarge) instance.
11. Start the instance.
12. `journalctl -b | grep "Swap already exists!"` to verify swap is created due to larger swap required.
Aug 04 18:25:03 ip-172-31-22-174.ec2.internal hibinit-agent[2600]: Swap already exists! (have 4194304000, need 8011620352), deleting existing swap file /swap
[Where things could go wrong]
- If Dynamic swap allocation fails, then hibernation may not properly save the state of the machine.
- Perhaps the user relies on the smaller/bigger SWAP size for something else and dynamically resizing on instance change could cause issues? I'd say unlikely but could be a concern. |
|
2023-08-30 22:31:33 |
Steve Langasek |
ec2-hibinit-agent (Ubuntu): status |
New |
Triaged |
|
2023-08-31 10:50:51 |
Launchpad Janitor |
ec2-hibinit-agent (Ubuntu): status |
Triaged |
Fix Released |
|
2023-08-31 14:35:39 |
Robie Basak |
nominated for series |
|
Ubuntu Jammy |
|
2023-08-31 14:35:39 |
Robie Basak |
bug task added |
|
ec2-hibinit-agent (Ubuntu Jammy) |
|
2023-08-31 14:35:39 |
Robie Basak |
nominated for series |
|
Ubuntu Lunar |
|
2023-08-31 14:35:39 |
Robie Basak |
bug task added |
|
ec2-hibinit-agent (Ubuntu Lunar) |
|
2023-08-31 14:35:39 |
Robie Basak |
nominated for series |
|
Ubuntu Focal |
|
2023-08-31 14:35:39 |
Robie Basak |
bug task added |
|
ec2-hibinit-agent (Ubuntu Focal) |
|
2023-09-06 00:41:25 |
Launchpad Janitor |
merge proposal linked |
|
https://code.launchpad.net/~mitchdz/ubuntu/+source/ec2-hibinit-agent/+git/ec2-hibinit-agent/+merge/450742 |
|
2023-09-06 14:26:25 |
Launchpad Janitor |
merge proposal linked |
|
https://code.launchpad.net/~mitchdz/ubuntu/+source/ec2-hibinit-agent/+git/ec2-hibinit-agent/+merge/450796 |
|
2023-09-06 14:26:44 |
Launchpad Janitor |
merge proposal linked |
|
https://code.launchpad.net/~mitchdz/ubuntu/+source/ec2-hibinit-agent/+git/ec2-hibinit-agent/+merge/450797 |
|
2023-09-06 15:43:01 |
Ubuntu Archive Robot |
bug |
|
|
added subscriber Graham Inggs |
2023-09-06 16:23:51 |
Graham Inggs |
ec2-hibinit-agent (Ubuntu Focal): assignee |
|
Graham Inggs (ginggs) |
|
2023-09-06 16:23:53 |
Graham Inggs |
ec2-hibinit-agent (Ubuntu Jammy): assignee |
|
Graham Inggs (ginggs) |
|
2023-09-06 16:23:57 |
Graham Inggs |
ec2-hibinit-agent (Ubuntu Lunar): assignee |
|
Graham Inggs (ginggs) |
|
2023-09-06 16:24:01 |
Graham Inggs |
ec2-hibinit-agent (Ubuntu Focal): status |
New |
In Progress |
|
2023-09-06 16:24:05 |
Graham Inggs |
ec2-hibinit-agent (Ubuntu Jammy): status |
New |
In Progress |
|
2023-09-06 16:24:08 |
Graham Inggs |
ec2-hibinit-agent (Ubuntu Lunar): status |
New |
In Progress |
|
2023-09-07 17:26:49 |
Mitchell Dzurick |
description |
Upstream Change - https://github.com/aws/amazon-ec2-hibinit-agent/commit/d2ce783d10d3a4437ebac468ed669666d8b6671b
[FFE]
====
This change includes a small bugfix to resize the swap partition. On hibernation, the contents of RAM is put onto SWAP, so if a user changes the amount of RAM on an instance, SWAP potentially needs to be recreated.
[SRU]
====
[Impact]
If an AWS EC2 user increases the amount of RAM on an instance, hibernation will fail.
This update dynamically allocates the SWAP to the necessary amount for the user. This has the added benefit of downsizing SWAP on RAM reduction thus optimizing disk space in those scenarios.
[Test Case]
1. create & start a c5.xlarge instance.
2. Upgrade to the new ec2-hibinit-agent package.
3. Stop the instance.
4. Modified to a smaller instance (c5.large)
5. `journalctl -b | grep "Swap already exists!"` verify swap is recreated
Aug 04 18:19:03 ip-172-31-22-174.ec2.internal hibinit-agent[2457]: Swap already exists! (have 8011616256, need 4194304000), deleting existing swap file /swap since current swap is sufficiently large and wasting memory
7. Stop and restart the instance again
8. `journalctl -b | grep "There's sufficient swap available"` to verified no recreation. of swap occured.
Aug 04 18:23:17 ip-172-31-22-174.ec2.internal hibinit-agent[2452]: There's sufficient swap available (have 4194304000, need 4194304000)
9. Stop the instance
10. Modify to larger (c5.xlarge) instance.
11. Start the instance.
12. `journalctl -b | grep "Swap already exists!"` to verify swap is created due to larger swap required.
Aug 04 18:25:03 ip-172-31-22-174.ec2.internal hibinit-agent[2600]: Swap already exists! (have 4194304000, need 8011620352), deleting existing swap file /swap
[Where things could go wrong]
- If Dynamic swap allocation fails, then hibernation may not properly save the state of the machine.
- Perhaps the user relies on the smaller/bigger SWAP size for something else and dynamically resizing on instance change could cause issues? I'd say unlikely but could be a concern. |
Upstream Change - https://github.com/aws/amazon-ec2-hibinit-agent/commit/d2ce783d10d3a4437ebac468ed669666d8b6671b
[FFE]
====
This change includes a small optimization to resize the swap partition. On hibernation, the contents of RAM is put onto SWAP, so if the amount of RAM reduces, the SWAP file will be too large, and waste space.
[SRU]
====
[Impact]
If an AWS EC2 user decreases the amount of RAM on an instance, unnecessary disk space will be used.
[Test Case]
1. create & start a c5.xlarge instance.
2. Upgrade to the new ec2-hibinit-agent package.
3. Stop the instance.
4. Modified to a smaller instance (c5.large)
5. `journalctl -b | grep "Swap already exists!"` verify swap is recreated
Aug 04 18:19:03 ip-172-31-22-174.ec2.internal hibinit-agent[2457]: Swap already exists! (have 8011616256, need 4194304000), deleting existing swap file /swap since current swap is sufficiently large and wasting memory
7. Stop and restart the instance again
8. `journalctl -b | grep "There's sufficient swap available"` to verified no recreation. of swap occured.
Aug 04 18:23:17 ip-172-31-22-174.ec2.internal hibinit-agent[2452]: There's sufficient swap available (have 4194304000, need 4194304000)
9. Stop the instance
10. Modify to larger (c5.xlarge) instance.
11. Start the instance.
12. `journalctl -b | grep "Swap already exists!"` to verify swap is created due to larger swap required.
Aug 04 18:25:03 ip-172-31-22-174.ec2.internal hibinit-agent[2600]: Swap already exists! (have 4194304000, need 8011620352), deleting existing swap file /swap
[Where things could go wrong]
- If Dynamic swap allocation fails, then hibernation may not properly save the state of the machine.
- Perhaps the user relies on the bigger SWAP size for something else and dynamically resizing on instance change could cause issues? I'd say unlikely but could be a concern. |
|
2023-09-07 18:24:37 |
Mitchell Dzurick |
summary |
FFE+SRU: Add recreating the swap file if the current size is sufficiently larger |
Swap file remains unnecessarily large if system memory size decreases |
|
2023-09-08 09:23:24 |
Robie Basak |
description |
Upstream Change - https://github.com/aws/amazon-ec2-hibinit-agent/commit/d2ce783d10d3a4437ebac468ed669666d8b6671b
[FFE]
====
This change includes a small optimization to resize the swap partition. On hibernation, the contents of RAM is put onto SWAP, so if the amount of RAM reduces, the SWAP file will be too large, and waste space.
[SRU]
====
[Impact]
If an AWS EC2 user decreases the amount of RAM on an instance, unnecessary disk space will be used.
[Test Case]
1. create & start a c5.xlarge instance.
2. Upgrade to the new ec2-hibinit-agent package.
3. Stop the instance.
4. Modified to a smaller instance (c5.large)
5. `journalctl -b | grep "Swap already exists!"` verify swap is recreated
Aug 04 18:19:03 ip-172-31-22-174.ec2.internal hibinit-agent[2457]: Swap already exists! (have 8011616256, need 4194304000), deleting existing swap file /swap since current swap is sufficiently large and wasting memory
7. Stop and restart the instance again
8. `journalctl -b | grep "There's sufficient swap available"` to verified no recreation. of swap occured.
Aug 04 18:23:17 ip-172-31-22-174.ec2.internal hibinit-agent[2452]: There's sufficient swap available (have 4194304000, need 4194304000)
9. Stop the instance
10. Modify to larger (c5.xlarge) instance.
11. Start the instance.
12. `journalctl -b | grep "Swap already exists!"` to verify swap is created due to larger swap required.
Aug 04 18:25:03 ip-172-31-22-174.ec2.internal hibinit-agent[2600]: Swap already exists! (have 4194304000, need 8011620352), deleting existing swap file /swap
[Where things could go wrong]
- If Dynamic swap allocation fails, then hibernation may not properly save the state of the machine.
- Perhaps the user relies on the bigger SWAP size for something else and dynamically resizing on instance change could cause issues? I'd say unlikely but could be a concern. |
Upstream Change - https://github.com/aws/amazon-ec2-hibinit-agent/commit/d2ce783d10d3a4437ebac468ed669666d8b6671b
[FFE]
====
This change includes a small optimization to resize the swap partition. On hibernation, the contents of RAM is put onto SWAP, so if the amount of RAM reduces, the SWAP file will be too large, and waste space.
[SRU]
====
[Impact]
If an AWS EC2 user decreases the amount of RAM on an instance, unnecessary disk space will be used.
[Test Case]
1. create & start a c5.xlarge instance.
2. Upgrade to the new ec2-hibinit-agent package.
3. Stop the instance.
4. Modified to a smaller instance (c5.large)
5. `journalctl -b | grep "Swap already exists!"` verify swap is recreated
Aug 04 18:19:03 ip-172-31-22-174.ec2.internal hibinit-agent[2457]: Swap already exists! (have 8011616256, need 4194304000), deleting existing swap file /swap since current swap is sufficiently large and wasting memory
7. Stop and restart the instance again
8. `journalctl -b | grep "There's sufficient swap available"` to verified no recreation. of swap occured.
Aug 04 18:23:17 ip-172-31-22-174.ec2.internal hibinit-agent[2452]: There's sufficient swap available (have 4194304000, need 4194304000)
9. Stop the instance
10. Modify to larger (c5.xlarge) instance.
11. Start the instance.
12. `journalctl -b | grep "Swap already exists!"` to verify swap is created due to larger swap required.
Aug 04 18:25:03 ip-172-31-22-174.ec2.internal hibinit-agent[2600]: Swap already exists! (have 4194304000, need 8011620352), deleting existing swap file /swap
[Where things could go wrong]
- If Dynamic swap allocation fails, then hibernation may not properly save the state of the machine.
- Perhaps the user relies on the bigger SWAP size for something else and dynamically resizing on instance change could cause issues? I'd say unlikely but could be a concern.
- [racb] I asked what might happen if a user has increased swap size manually and expects it to stay large. But I'm told that a specific swap filename is used that is specific to hibernation, so it seems reasonable that this would be automatically managed. |
|
2023-09-08 09:26:09 |
Robie Basak |
ec2-hibinit-agent (Ubuntu Lunar): status |
In Progress |
Fix Committed |
|
2023-09-08 09:26:11 |
Robie Basak |
bug |
|
|
added subscriber Ubuntu Stable Release Updates Team |
2023-09-08 09:26:12 |
Robie Basak |
bug |
|
|
added subscriber SRU Verification |
2023-09-08 09:26:15 |
Robie Basak |
tags |
|
verification-needed verification-needed-lunar |
|
2023-09-08 09:26:25 |
Robie Basak |
ec2-hibinit-agent (Ubuntu Jammy): status |
In Progress |
Fix Committed |
|
2023-09-08 09:26:28 |
Robie Basak |
tags |
verification-needed verification-needed-lunar |
verification-needed verification-needed-jammy verification-needed-lunar |
|
2023-09-08 09:26:37 |
Robie Basak |
ec2-hibinit-agent (Ubuntu Focal): status |
In Progress |
Fix Committed |
|
2023-09-08 09:26:40 |
Robie Basak |
tags |
verification-needed verification-needed-jammy verification-needed-lunar |
verification-needed verification-needed-focal verification-needed-jammy verification-needed-lunar |
|
2023-09-25 22:48:22 |
Mitchell Dzurick |
tags |
verification-needed verification-needed-focal verification-needed-jammy verification-needed-lunar |
verification-done verification-done-focal verification-done-jammy verification-done-lunar |
|
2023-09-26 22:20:33 |
Launchpad Janitor |
ec2-hibinit-agent (Ubuntu Lunar): status |
Fix Committed |
Fix Released |
|
2023-09-26 22:20:38 |
Brian Murray |
removed subscriber Ubuntu Stable Release Updates Team |
|
|
|
2023-09-26 22:21:13 |
Launchpad Janitor |
ec2-hibinit-agent (Ubuntu Jammy): status |
Fix Committed |
Fix Released |
|
2023-09-26 22:21:50 |
Launchpad Janitor |
ec2-hibinit-agent (Ubuntu Focal): status |
Fix Committed |
Fix Released |
|
2023-09-26 22:30:27 |
Brian Murray |
bug |
|
|
added subscriber Brian Murray |