Commit Graph

5837 Commits

Author SHA1 Message Date
Adrien Delorme aae1992649 remove default PauseBeforeSSM, this will have to be set manually 2020-10-29 13:38:03 +01:00
Adrien Delorme 6c45f04467 Delete ssm_mock_funcs.go 2020-10-29 13:37:44 +01:00
Adrien Delorme 5d06a6e6df rename file correctly 2020-10-29 13:26:09 +01:00
Adrien Delorme c6e2dd5538 remove unit test file for now, I think that an acceptance test will be easier here 2020-10-29 13:22:17 +01:00
Adrien Delorme a4bd744955 simplify things a bit more 2020-10-29 13:11:07 +01:00
Adrien Delorme aef3d24213 Update step_create_ssm_tunnel.go 2020-10-29 12:31:01 +01:00
Adrien Delorme f329cb5b93 simplify code 2020-10-29 12:18:41 +01:00
Adrien Delorme b058de072a move packer/builder/amazon/common.IsAWSErr to builder/amazon/common/awserrors.Matches
to avoid cyclic dependency issues
2020-10-29 12:02:41 +01:00
Adrien Delorme 8e355d0fe7 Move ssm code to its own ssm package and make it singlethreaded 2020-10-29 11:57:29 +01:00
Wilken Rivera 646b973bd3 Remove logic to retry a connection that reuses an existing SSM Session
After testing it was found that once an session is terminated via an
instance restart, console termination, or SSM agent restart. Any active
session will essentially be terminated and unusable. So knowing that it
is always best to start a new session and let the old one timeout get
terminated.
2020-10-28 06:36:23 -04:00
Wilken Rivera dff9cde775 Remove waitgroups 2020-10-28 06:36:23 -04:00
Wilken Rivera eb11009e2a Check for closed channels as opposed to using a separate closeRetry channel 2020-10-28 06:36:23 -04:00
sylviamoss aa73cc7d7e add close chan to avoid unwanted retries 2020-10-27 16:32:20 -04:00
sylviamoss 1f62249097 add retry terminated session chan 2020-10-27 16:32:20 -04:00
sylviamoss 8e3f3e514c improve logs 2020-10-27 16:32:20 -04:00
sylviamoss b2c7897f58 add WaitGroup to avoid data race 2020-10-27 16:32:20 -04:00
sylviamoss 1c2b469acd add retry channel to ssm driver 2020-10-27 16:32:19 -04:00
blz-ea 73370570f4
Fix Proxmox builder unhandled buildvar type (#10154)
* Fix Proxmox builder unhandled buildvar type #10139

Co-authored-by: Calle Pettersson <carlpett@users.noreply.github.com>
Co-authored-by: Adrien Delorme <adrien.delorme@icloud.com>
2020-10-26 15:11:09 +01:00
Megan Marsh 25f4e24772
Merge pull request #10150 from nshalman/nshalman/tag-snapshots
amazon/ebssurrogate: apply snapshot tags right when taking snapshot
2020-10-23 15:57:18 -07:00
Megan Marsh 4862b2c0f0
Merge pull request #10163 from hashicorp/fix-azure_arm-manageddisk-deletion-regression
builder/azure_arm: Fix build failures due to the deletion of attached managed disks
2020-10-23 15:19:30 -07:00
Megan Marsh fc619dc977
Merge pull request #9626 from featheredtoast/add-proxmox-vm-clone
builder/proxmox FEATURE: split Proxmox into proxmox-iso and proxmox-clone
2020-10-23 13:36:45 -07:00
Megan Marsh 56c7a9cda5 rebase 2020-10-23 11:47:20 -07:00
netapp-jgriffit 15467aa868 Update builder/vsphere/common/step_shutdown.go
Co-authored-by: Sylvia Moss <sylviamoss.m@gmail.com>
2020-10-23 11:41:03 -07:00
James Griffith 185f3d9d48 vsphere clone config not yet using warnings, leaving code commented out 2020-10-23 11:41:03 -07:00
James Griffith c8ad66f419 update to docs as suggested 2020-10-23 11:41:03 -07:00
James Griffith c86a48fbc9 syntax change 2020-10-23 11:41:03 -07:00
James Griffith f86f7c4464 Another instance using `Prepare()` needed an update 2020-10-23 11:41:03 -07:00
James Griffith 09edada93d moved the warning into Prepare() and updated the invocation 2020-10-23 11:41:03 -07:00
James Griffith 572de129d7 update documentation 2020-10-23 11:41:03 -07:00
James Griffith 1048cdbcc4 check for no communicator and provide messaging to user 2020-10-23 11:41:03 -07:00
James Griffith 2ecf5cc9aa Communicator is not needed unless shutdown_command is populated 2020-10-23 11:41:03 -07:00
James Griffith 6b39b1eed6 StepShutdown should still occur if Comm.Type is none 2020-10-23 11:41:03 -07:00
James Griffith 65ff092fd4 hashicorp/packer#9924 2020-10-23 11:41:03 -07:00
Wilken Rivera a1d69bfe91 Fix liniting issues 2020-10-23 13:52:05 -04:00
Wilken Rivera afd33679f5 builder/azure_arm: Fix build failures due to the deletion of attached managed disks
Previously (prior to v1.6.2) the Azure ARM builder had two delete functions
one that would run before any of the StepDelete* types, and one on deployment template
cleanup. The refactored coded re-introduces the logic from the previously removed
step in v1.6.1 as the main delete logic for the whole deployment
template. Ensuring that all deployed items are deleted before trying to
remove any managed disks.

This change moves all the deletion logic into the
step_deployment_template#Cleanup function to ensure that dependent steps
are only called once the created deployment items (i.e
StepDelateAdditionalDisks) have been deleted.

Test results before change
```

    compute.DisksClient#Delete: Failure sending request: StatusCode=0 -- Original Error: autorest/azure: Service returned an error. Status=<nil> Code="OperationNotAllowed" Message="Disk pkrdd02e9rzzu5k-1 is attached to VM /subscriptions/1f90521a-24f6-4758-ac3d-88d869fb0bf5/resourceGroups/packer-acceptance-test/providers/Microsoft.Compute/virtualMachines/pkrvm02e9rzzu5k."
--- FAIL: TestBuilderAcc_ManagedDisk_Windows_Build_Resource_Group_Additional_Disk (454.00s)
FAIL
FAIL    github.com/hashicorp/packer/builder/azure/arm   454.008s

```

Test results after change
```
--- PASS: TestBuilderAcc_ManagedDisk_Windows_Build_Resource_Group_Additional_Disk (563.56s)

```

Closes #10070
2020-10-23 13:44:27 -04:00
Megan Marsh 8f67f939f9
Merge pull request #10158 from Direnol/yandex-add-support-iam-auth
Added support for IAM credential in the token field and YC_TOKEN env
2020-10-23 09:39:46 -07:00
Roman Mingazeev 75befba67e add more doc 2020-10-23 18:01:52 +03:00
Roman Mingazeev 60595ad522 remove extra log 2020-10-23 14:09:05 +03:00
Roman Mingazeev 72fe3bbc26
Update builder/yandex/access_config.go
Co-authored-by: GennadySpb <lipenkov@gmail.com>
2020-10-23 14:05:39 +03:00
Roman Mingazeev ae7fc84557 Added support for IAM credential in the token field and YC_TOKEN env 2020-10-23 11:47:28 +03:00
Megan Marsh 3be53e10af
Merge pull request #10128 from nayyara-samuel/add-state-timeout-linode
Add documented `state_timeout` parameters to Linode builder
2020-10-22 15:13:59 -07:00
Megan Marsh c4001734d0
Merge pull request #10143 from hashicorp/do_9951
builder/vsphere: skip iso download if hashed file is already present on remote datastore
2020-10-22 15:06:04 -07:00
Megan Marsh 00037d8e69
Merge pull request #10145 from hashicorp/fix_10135
commhostport isn't present if communicator is null; protect casting i…
2020-10-22 14:24:23 -07:00
Megan Marsh 70cb4097e1 review comments, add plugin type for a few missing builders 2020-10-21 16:17:32 -07:00
Megan Marsh c0ce8a9414 rework fixer deprecation code so we know what plugins they relate to 2020-10-21 15:28:59 -07:00
Megan Marsh f1f834b8fe commhostport isn't present if communicator is null; protect casting in a check 2020-10-21 14:30:23 -07:00
Nahum Shalman 83a672f2c9 amazon/ebssurrogate: apply snapshot tags right when taking snapshot 2020-10-21 16:17:37 -04:00
Megan Marsh 796c40f89b builder/vsphere: skip iso download if hashed file is already present in remote packer_cache 2020-10-21 11:33:43 -07:00
Wilken Rivera a2ba700ac0 Remove deprecated object ids from configs 2020-10-21 13:56:22 -04:00
Wilken Rivera 1621ab59d8
tests/azure/arm: Update test image name (#10133)
I've run into a few cases where running the full test suite for the ARM
builder would fail because of conflicting image names. This is a
workaround for now the bigger fix is to have the acceptance clean up any
created images.

Results before change
```
the managed image named testBuilderAccManagedDiskWindows-1603151855 already exists in the resource group packer-acceptance-test, use the -force option to automatically delete it.
--- FAIL: TestBuilderAcc_ManagedDisk_Windows_Build_Resource_Group (0.66s)
```

Results after change
```
PASS
ok      github.com/hashicorp/packer/builder/azure/arm   2431.115s

```
2020-10-21 11:12:50 +02:00