Commit Graph

162 Commits

Author SHA1 Message Date
14c7d5e4d9 mark dependant tasks failed upon failure (#650)
Fix #644
2021-03-11 22:24:43 +00:00
b4ceb263e0 stop jobs once all tasks are stopped (#649)
Fixed #643
2021-03-09 20:09:18 +00:00
0a3812d8bc Add job stopped task info (#648) 2021-03-09 10:06:06 -05:00
23dc8ad301 explain the source of task failures related notifications (#635) 2021-03-06 13:35:09 +00:00
92c1d0a7a1 only set VM passwords on Windows (#620) 2021-03-03 21:27:56 +00:00
78d6adf555 upgrade azure-storage-blob to 12.8.0 (#625)
Note, this makes use of the new feature, `ContainerClient.exists()` which models our existing mechanism for doing container existence checking.
2021-03-03 10:33:23 +00:00
04fc41597e require target_exe to be a canonicalized relative path (#613) 2021-03-02 19:06:02 -05:00
4489036d9f add node & task heartbeat events (#621)
This adds node & task heartbeats and makes the event data available as a structured data in the logs.
2021-03-02 22:04:39 +00:00
7f66eeee0d handle OperationNotAllowed errors when creating VMSS (#614) 2021-03-02 16:14:10 -05:00
c537458ade update azure-mgmt-compute to 19.0.0 (#611) 2021-03-02 18:44:26 +00:00
296ba2ee23 update azure-mgmt-storage to 17.0.0 (#612) 2021-03-02 18:00:07 +00:00
32681b2611 update azure-mgmt-resource (#607) 2021-03-02 08:35:07 +00:00
06f45f338c Update Task Heartbeat to include Job_id (#594) 2021-02-26 13:36:10 -05:00
8600a44f1f fix bool queries (#597)
This addresses broken queries used for identifying outdated nodes.
2021-02-26 16:51:05 +00:00
daef1637f8 update jinja2 (#595) 2021-02-26 09:19:10 -05:00
fb482e357e don't schedule work to a node if the scaleset or pool is shutting down (#583) 2021-02-23 13:33:41 -05:00
cebb84b9e7 handle error condition when creating a container that is being deleted (#582)
When users try to create a container immediately after deleting it, Azure will fail saying the deletion is in-progress.

catching ResourceExistsError during create handles this error.
2021-02-22 01:49:07 +00:00
feb80ecb54 allow nodes with multiple tasks to continue on task stop (#567)
As is, when multiple tasks are running on a single node, if any one of them stops, the node gets reimaged.

This changes the behavior such that when a node with multiple tasks has one task stop, the other tasks will continue.
2021-02-19 23:54:26 +00:00
6ba5795f36 update proxy port ranges to avoid current blocks (#552) 2021-02-19 17:50:09 -05:00
4de19ffe5e stop jobs that do not start within 30 days (#565)
If a job does not start within 30 days, stop the job and mark all of the tasks as `failed`.
2021-02-19 21:23:35 +00:00
305c23a4d9 add instance information to webhooks (#577)
Fixes #574
2021-02-19 21:00:51 +00:00
8ce4638b8a clarify scaleset logging (#568) 2021-02-19 19:36:16 +00:00
4992b494f1 add task config to all task events (#580) 2021-02-19 14:10:48 -05:00
872a5ddc14 add details to exceptions generated during report render failures (#576) 2021-02-19 13:48:49 -05:00
3a7bc95316 import local relative paths (#579) 2021-02-19 12:29:35 -05:00
929d9ce496 make user triggered reimaging happen immediately (#566) 2021-02-18 14:08:25 -05:00
279629292f handle SkuNotAvailable errors when creating VM Scalesets (#557) 2021-02-17 16:52:37 -05:00
bdcab6eb08 handle tokens from x-ms-token-aad-id-token (#531) 2021-02-10 12:41:15 -05:00
8ee7fae240 use the cached Azure Identity instance for storage operations (#526) 2021-02-09 12:20:12 -05:00
91a3690551 fix logging to show corpus accounts found, not a function ref (#524) 2021-02-09 11:55:52 -05:00
5114332ea0 clarify proxy log messages (#520) 2021-02-08 17:39:26 -05:00
8c9f65c0be add missing scaleset nodes (#518) 2021-02-08 13:50:08 -05:00
fd995718c8 standardize on unix file endings (#516)
The majority of our source files have `\n` line endings.  This updates the few files that use `\r\n` to use `\n`.
2021-02-06 14:14:26 +00:00
1d74379a70 use the primitive types in more places (#514) 2021-02-05 13:10:37 -05:00
51f4eea069 update nodes and scalesets in a consistent order (#512) 2021-02-05 12:51:38 -05:00
e3dfcb8b95 Scalesets that are about to be deleted don't need updated configs (#511) 2021-02-05 09:53:29 -05:00
3cb055d331 clarify message upon service & agent version mismatch (#510) 2021-02-04 19:58:45 -05:00
a02e084522 split out node, scaleset, and pool code (#507) 2021-02-04 19:07:49 -05:00
c5bb0f0588 Update Proxy heartbeat & logging (#502) 2021-02-04 15:38:17 -05:00
1adee4ea83 update pools before nodes or scalesets (#503) 2021-02-04 13:26:19 -05:00
e0cd3d535b rename proxy queue endpoint (#504) 2021-02-04 12:45:36 -05:00
0a4278110d update azure-mgmt-compute to 18.2.0 (#499) 2021-02-04 12:34:29 +00:00
5e2e9448df add security auditing of python code using Bandit during CICD (#491) 2021-02-01 16:51:03 -05:00
0f70ffa3e2 try pushing updates to scaleset configs frequently until the push succeeds (#489) 2021-02-01 10:09:40 -05:00
a46f7b4193 expose supervisor tasks that are fully self-contained fuzzing tasks in the service (#474)
Exposes the functionality added in #454 to the service & CLI.

Fixes #439
2021-01-29 00:01:59 +00:00
14fc1ca51f remove unused Event generation from the pre-2.0.0 SignalR integration (#477)
Remove a vestige of the adhoc events used by the previous SingalR integration for container updates.
2021-01-28 21:56:31 +00:00
f155ad625f reimage long-lived nodes (#476)
This helps keep nodes on scalesets that use `latest` OS image SKUs reasonably up-to-date with OS patches without disrupting running fuzzing tasks with patch reboot cycles.

In combination with the already-merged #416, this PR closes #414.
2021-01-28 20:36:40 +00:00
2e44a5a22f update azure-mgmt-network to the latest version (#471) 2021-01-28 07:04:52 +00:00
24685ca8df Updating Windows Default Image from RS5-Pro to 20H2-Pro (#469)
RS5-Pro is no longer updated in the Azure Marketplace. In order to ensure the Windows 10 VMs are regularly updated, we need to switch the default image to 20H2-Pro, which is regularly maintained.
2021-01-27 13:46:46 +00:00
5027745ee2 simplify get/delete for scalesets (#468) 2021-01-26 14:43:14 -05:00