Fix: Remove container directory cleanup on exec process start failures

[Constantine3]

When container exec fails (e.g., invalid path), the container directory is deleted but the container remains in the in-memory list. Subsequent container rm fails because the config.json file no longer exists.

Root cause: The cleanup logic in createProcess() and startProcess() didn't distinguish between init process failures and exec process failures.

Solution:

1. Removed cleanup from createProcess() - if process creation fails, the container directory should remain since no state has been modified.
2. Made cleanup in startProcess() conditional using the existing isInitProcess() helper - cleanup only triggers when id == processID (init process), keeping the container intact on exec process failures.

Testing(Manual verification):

1. Created container fixtest2 with run --name fixtest2 oraclelinux:10 /bin/sleep 1000
2. Executed invalid command: exec fixtest2 /nonexistent/path - error returned
3. Verified container directory still exists at containers/fixtest2/
4. Validated container rm fixtest2 succeeds

Motivation and Context:
Fixes#1058 - Unable to delete container after exec failure

[jglogan]

@JaewonHur Could you look at this one and see how it relates to your work in #1080?

[Constantine3]

When container exec fails (e.g., invalid path), the container directory is deleted but the container remains in the in-memory list. Subsequent container rm fails because the config.json file no longer exists.

Root cause: The cleanup logic in createProcess() and startProcess() didn't distinguish between init process failures and exec process failures.

Solution:

1. Removed cleanup from createProcess() - if process creation fails, the container directory should remain since no state has been modified.
2. Made cleanup in startProcess() conditional using the existing isInitProcess() helper - cleanup only triggers when id == processID (init process), keeping the container intact on exec process failures.

Testing(Manual verification):

1. Created container fixtest2 with run --name fixtest2 oraclelinux:10 /bin/sleep 1000
2. Executed invalid command: exec fixtest2 /nonexistent/path - error returned
3. Verified container directory still exists at containers/fixtest2/
4. Validated container rm fixtest2 succeeds

Motivation and Context:
Fixes#1058 - Unable to delete container after exec failure

[jglogan]

@JaewonHur Could you look at this one and see how it relates to your work in #1080?

[Constantine3]

When container exec fails (e.g., invalid path), the container directory is deleted but the container remains in the in-memory list. Subsequent container rm fails because the config.json file no longer exists.

Root cause: The cleanup logic in createProcess() and startProcess() didn't distinguish between init process failures and exec process failures.

Solution:

1. Removed cleanup from createProcess() - if process creation fails, the container directory should remain since no state has been modified.
2. Made cleanup in startProcess() conditional using the existing isInitProcess() helper - cleanup only triggers when id == processID (init process), keeping the container intact on exec process failures.

Testing(Manual verification):

1. Created container fixtest2 with run --name fixtest2 oraclelinux:10 /bin/sleep 1000
2. Executed invalid command: exec fixtest2 /nonexistent/path - error returned
3. Verified container directory still exists at containers/fixtest2/
4. Validated container rm fixtest2 succeeds

Motivation and Context:
Fixes#1058 - Unable to delete container after exec failure

[jglogan]

@JaewonHur Could you look at this one and see how it relates to your work in #1080?

[Constantine3]

When container exec fails (e.g., invalid path), the container directory is deleted but the container remains in the in-memory list. Subsequent container rm fails because the config.json file no longer exists.

Root cause: The cleanup logic in createProcess() and startProcess() didn't distinguish between init process failures and exec process failures.

Solution:

1. Removed cleanup from createProcess() - if process creation fails, the container directory should remain since no state has been modified.
2. Made cleanup in startProcess() conditional using the existing isInitProcess() helper - cleanup only triggers when id == processID (init process), keeping the container intact on exec process failures.

Testing(Manual verification):

1. Created container fixtest2 with run --name fixtest2 oraclelinux:10 /bin/sleep 1000
2. Executed invalid command: exec fixtest2 /nonexistent/path - error returned
3. Verified container directory still exists at containers/fixtest2/
4. Validated container rm fixtest2 succeeds

Motivation and Context:
Fixes#1058 - Unable to delete container after exec failure

[jglogan]

@JaewonHur Could you look at this one and see how it relates to your work in #1080?

[Constantine3]

When container exec fails (e.g., invalid path), the container directory is deleted but the container remains in the in-memory list. Subsequent container rm fails because the config.json file no longer exists.

Root cause: The cleanup logic in createProcess() and startProcess() didn't distinguish between init process failures and exec process failures.

Solution:

1. Removed cleanup from createProcess() - if process creation fails, the container directory should remain since no state has been modified.
2. Made cleanup in startProcess() conditional using the existing isInitProcess() helper - cleanup only triggers when id == processID (init process), keeping the container intact on exec process failures.

Testing(Manual verification):

1. Created container fixtest2 with run --name fixtest2 oraclelinux:10 /bin/sleep 1000
2. Executed invalid command: exec fixtest2 /nonexistent/path - error returned
3. Verified container directory still exists at containers/fixtest2/
4. Validated container rm fixtest2 succeeds

Motivation and Context:
Fixes#1058 - Unable to delete container after exec failure

[jglogan]

@JaewonHur Could you look at this one and see how it relates to your work in #1080?

[Constantine3]

When container exec fails (e.g., invalid path), the container directory is deleted but the container remains in the in-memory list. Subsequent container rm fails because the config.json file no longer exists.

Root cause: The cleanup logic in createProcess() and startProcess() didn't distinguish between init process failures and exec process failures.

Solution:

1. Removed cleanup from createProcess() - if process creation fails, the container directory should remain since no state has been modified.
2. Made cleanup in startProcess() conditional using the existing isInitProcess() helper - cleanup only triggers when id == processID (init process), keeping the container intact on exec process failures.

Testing(Manual verification):

1. Created container fixtest2 with run --name fixtest2 oraclelinux:10 /bin/sleep 1000
2. Executed invalid command: exec fixtest2 /nonexistent/path - error returned
3. Verified container directory still exists at containers/fixtest2/
4. Validated container rm fixtest2 succeeds

Motivation and Context:
Fixes#1058 - Unable to delete container after exec failure

[jglogan]

@JaewonHur Could you look at this one and see how it relates to your work in #1080?

[Constantine3]

When container exec fails (e.g., invalid path), the container directory is deleted but the container remains in the in-memory list. Subsequent container rm fails because the config.json file no longer exists.

Root cause: The cleanup logic in createProcess() and startProcess() didn't distinguish between init process failures and exec process failures.

Solution:

1. Removed cleanup from createProcess() - if process creation fails, the container directory should remain since no state has been modified.
2. Made cleanup in startProcess() conditional using the existing isInitProcess() helper - cleanup only triggers when id == processID (init process), keeping the container intact on exec process failures.

Testing(Manual verification):

1. Created container fixtest2 with run --name fixtest2 oraclelinux:10 /bin/sleep 1000
2. Executed invalid command: exec fixtest2 /nonexistent/path - error returned
3. Verified container directory still exists at containers/fixtest2/
4. Validated container rm fixtest2 succeeds

Motivation and Context:
Fixes#1058 - Unable to delete container after exec failure

[jglogan]

@JaewonHur Could you look at this one and see how it relates to your work in #1080?

[Constantine3]

When container exec fails (e.g., invalid path), the container directory is deleted but the container remains in the in-memory list. Subsequent container rm fails because the config.json file no longer exists.

Root cause: The cleanup logic in createProcess() and startProcess() didn't distinguish between init process failures and exec process failures.

Solution:

1. Removed cleanup from createProcess() - if process creation fails, the container directory should remain since no state has been modified.
2. Made cleanup in startProcess() conditional using the existing isInitProcess() helper - cleanup only triggers when id == processID (init process), keeping the container intact on exec process failures.

Testing(Manual verification):

1. Created container fixtest2 with run --name fixtest2 oraclelinux:10 /bin/sleep 1000
2. Executed invalid command: exec fixtest2 /nonexistent/path - error returned
3. Verified container directory still exists at containers/fixtest2/
4. Validated container rm fixtest2 succeeds

Motivation and Context:
Fixes#1058 - Unable to delete container after exec failure

[jglogan]

@JaewonHur Could you look at this one and see how it relates to your work in #1080?