-
Notifications
You must be signed in to change notification settings - Fork 21
/
RELEASE_NOTES.txt
179 lines (134 loc) · 4.85 KB
/
RELEASE_NOTES.txt
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
# v0.9.68
- bug fixes
- remove some depricated code
- fix /queue resource
- fix task reset
- add recoverall option for admin use
- update IO functions, reduce unecessary shock GETs
- handle asynchronous shock actions
- node file and index locks
- wait on unlock (with timeout) during IO validation
# v0.9.67
- more CWL support
- bug fixes
- improved error handling / messages
- fix worker errors being ignored
- parts node index triggered by server and not worker
- reset task on: resume, recompute, resubmit
- locking when reading / modifying task.IO
- bookkeeping / sync: job state, taskmap, node url, task dependants
# v0.9.51
- add backwords compatible use of 'oauth' bearer key when using single auth
# v0.9.50
- fix workunit.ExitStatus on completion
- add coReq channel length as a config option, default 100
- fix work discarded state
# v0.9.49
- support multiple oauth servers
- requires config file / cli option change
- template config files updated
- auth error handling fixes
- buffer work request channel
- more bug fixes
- update base url
# v0.9.48
- golang 1.7.6
- more cleanup and optimize job and task set functions
- overhaul / locking on initialization of task.Partition (multi-workunits)
- fix worker download of parts Inputs
- cache shock node indexinfo with io struct
- handle slow processing of completed work
- error handling updates
# v0.9.47
- new state FAILED_PERMANENT introduced
- awe-worker can execute cwl-runner
- added test for special exit code 42
- remove deprecated code
- cleanup and optimize job and task set functions
- add job.Error struct
- fix job suspension logic
# v0.9.46
- update shock struct
- fix missing output file check
# v0.9.45
- golang 1.7.4
- more bug fixes
- now: awe-worker
- alpha cwl support
# v0.9.43
- lots of bug fixes
- lots of locking
- globus auth fix
# v0.9.33
- fix mongo reading
- added metadata options for client
# v0.9.32
- fine-grained mongo updates instead of complete job writes
- completed fix of locking mechanism
# v0.9.31
- golang 1.7.3
- UI site update
- added RWLock to Client
# v0.9.30
- golang 1.7.1
- removed duplicate uuid vendor library
- more GET /client query options
- /client?group=<clientgroup name>
- /client?status=<client status>
- /client?app=<app in list>
- added GET /queue?clientgroup
- return list of running jobs for that clientgroup
- added /logger resource
- /logger?debug=N - PUT debug level
- /logger?debug - GET debug level
- /logger?event - GET event code descriptions
- removed unneeded / unused jid from job document and codebase (servermgr handle)
- workunit log cleanup - delete on task reset
- add server and client config option '--logoutput' to set output logs to file or console, or both
# v0.9.29
- vendor path changes for golang 1.6 support
- docker file updates
- alpha base image
- golang 1.6.3
# v0.9.27
- improved log / error messages
- added 'report' option to job GET, returns a summary of job with stdout, stderr, and worknotes for each workunit ran
- changed max jid to be stored in mongodb and not on filesystem
- this is backwards compatible, will use jid on filesystem for first time if no value is in mongodb
# v0.9.26
- added 'full' option to job DELETE, actually deletes job from mongoDB and filesystem
- added expiration to job document
- set globally or per pipeline in config (off by default)
- user settable
- time is from completion of job
- added 'resubmit' option to job PUT, re-starts job from beginning
- refactor re-enqueue functions: resume, recompute, resubmit, recover
- re-enqueueing a task now deletes the shock nodes of the outputs, no more orphaned nodes
- fix race-condition resulting in broken re-enqueueing of tasks
- add 'distinct' option to job query for list of distinct values, must be indexed field in job.info
- update base API response
- added /event resource - event code descriptions
- expanded /queue resource
- option for json output format
- admin user only options to view current state of in-memory queues
- added 'import' field to job POST, allow import of completed job documents into mongoDB
- add validation for some config parameters
# v0.9.25
- fix bug in server when checking out workunit
# v0.9.24
- explicit check for bson document max size upon save, return meaningful error
- test mongo connection with short timeout on startup
- client sends available disk size in bytes when requesting a workunit
- server filters workunits based on sum of input sizes compared to available disk space on client
# v0.9.23
- added suspend / resume action for server work queue, admin users only
# v0.9.22
- fixed bug with docker predata dir
- added config options to client that were previously hardcoded
# v0.9.21
- adding logging and code cleanup to App usage
- updates to dockerfile
- bug fixes
# v0.9.20
- Update to newer version of mgo (v2)
- Removed dockerized version of mgo lib that is no longer in use.