5 - Make parallel-find.pl use `sudo -u username git describe --tags`
7 - Make the installer use 'wizard install' /or/ do a migration
8 after doing a normal install (the latter makes it easier
11 - Better error message if daemon/scripts-security-upd
12 is not on scripts-security-upd list
14 - Fix retarded logging mechanism
15 - This is interesting because the non-retarded way would be
16 to have each subprocess send data through stdout, but
17 this runs the risk of filling up the tubes. Maybe just
18 create a log directory and give each process different
21 - The great initial deploy:
22 - Turn on mediawiki new autoinstaller
23 - Migrate all mediawiki installs
24 - Build research scripts, determine if a re-migration is necessary
26 - Make upgrade and install take version as a parameter
28 - Redo Wordpress conversion, with an eye for automating everything
29 possible (such as downloading the tarball and unpacking)
31 - Summary script should be more machine friendly, and should not
32 output summary charts when I increase specificity
34 Some other stuff to do in your copious free time:
35 - Summary script should do something intelligent when distinguishing
36 between old-style and new-style installs
37 - Check how many autoinstalls are missing w bits for
38 daemon.scripts (this would need pyafs)
39 - Make scripts AFS patch advertise its existence so we can check for it.
40 (This might be otherwise possible using `fs sysname`)
41 - Make 'wizard summary' generate nice pretty graphs of installs by date
42 (more histograms, will need to check actual .scripts-version files.)
43 - It should be able to handle installs like Django where there's a component
44 that gets installed in web_scripts and another directory that gets installed
47 PULLING OUT CONFIGURATION FILES IN AN AUTOMATED MANNER
49 advancedpoll: Template file to fill out
50 django: Noodles of template files
51 gallery2: Multistage install process
53 mediawiki: One-step install process
54 phpbb: Multistage install process
55 phpical: Template file
58 wordpress: Multistage install process
60 PHILOSOPHY ABOUT LOGGING
62 Logging is most useful when performing a mass run. This
63 includes things such as mass-migration as well as when running
64 summary reports. An interesting property about mass-migration
65 or mass-upgrade, however, is that if they fail, they are
66 idempotent, so an individual case can be debugged simply running
67 the single-install equivalent with --debug on. (This, indeed,
68 may be easier to do than sifting through a logfile).
70 It is a different story when you are running a summary report:
71 you are primarily bound by your AFS cache and how quickly you can
72 iterate through all of the autoinstalls. Checking if a file
73 exists on a cold AFS cache may
74 take several minutes to perform; on a hot cache the same report
75 may take a mere 3 seconds. When you get to more computationally
76 expensive calculations, however, even having a hot AFS cache
77 is not enough to cut down your runtime.
79 There are certain calculations that someone may want to be
80 able to perform on manipulated data. As such, this data should
81 be cached on disk, if the process for extracting this data takes
82 a long time. Also, for usability sake, Wizard should generate
83 the common case reports.
85 Ensuring that machine parseable reports are made, and then making
86 the machinery to reframe this data, increases complexity. Therefore,
87 the recommendation is to assume that if you need to run iteratively,
88 you'll have a hot AFS cache at your fingerprints, and if that's not
89 fast enough, then cache the data.
91 COMMIT MESSAGE FIELDS:
93 Installed-by: username@hostname
94 Pre-commit-by: Real Name <username@mit.edu>
95 Upgraded-by: Real Name <username@mit.edu>
96 Migrated-by: Real Name <username@mit.edu>
97 Wizard-revision: abcdef1234567890
98 Wizard-args: /wizard/bin/wizard foo bar baz
102 Committer: Real Name <username@mit.edu>
103 Author: lockername locker <lockername@scripts.mit.edu>
107 - It is not expected or required for update scripts to exist for all
108 intervening versions that were present pre-migration; only for it
109 to work on the most recent migration.
111 - Currently all repositories are initialized with --shared, which
112 means they have basically ~no space footprint. However, it
113 also means that /mit/scripts/wizard/srv MUST NOT lose revs after
116 - Full fledged logging options. Namely:
117 x all loggers (delay implementing this until we actually have debug stmts)
119 - debug => loglevel = DEBUG
121 - default is WARNING (see below for exception)
122 - verbose => loglevel = INFO
123 x file logger (creates a dir and lots of little logfiles)
125 - log-file => loglevel = INFO
129 * Some parts of the infrastructure will not be touched, although I plan
130 on documenting them. Specifically, we will be keeping:
132 - parallel-find.pl, and the resulting
133 /mit/scripts/.htaccess/scripts/sec-tools/store/scriptslist
135 * The new procedure for generating an update is as follows:
136 (check out the mass-migration instructions for something in this spirit,
137 although uglier in some ways; A indicates the step /should/ be automated)
139 0. ssh into not-backward, temporarily give the daemon.scripts-security-upd
140 bits by blanching it on system:scripts-security-upd, and run parallel-find.pl
142 1. Have the Git repository and working copy for the project on hand.
144 /- wizard prepare-pristine --
146 A 2. Checkout the pristine branch
148 A 3. Remove all files from the working copy. Use `wipe-working-dir`
150 A 4. Download the new tarball
152 A 5. Extract the tarball over the working copy (`cp -R a/. b` works well,
153 remember that the working copy is empty; this needs some intelligent
156 A 6. Check for empty directories and add stub files as necessary.
157 Use `preserve-empty-dir`
161 7. Git add it all, and then commit as a new pristine version (v1.2.3)
163 8. Checkout the master branch
165 9. [FOR EXISTING REPOSITORIES]
166 Merge the pristine branch in. Resolve any conflicts that our
167 patches have with new changes. Do NOT let Git auto-commit it
168 with --no-commit (otherwise, you want to git commit --amend
169 to keep our history clean
171 [FOR NEW REPOSITORIES]
172 Check if any patches are needed to make the application work
173 on Scripts (ideally, it shouldn't.
175 /- wizard prepare-new --
178 A echo "Deny from all" > .scripts/.htaccess
182 10. Check if there are any special update procedures, and update
183 the wizard.app.APPNAME module accordingly (or create it, if
186 11. Run 'wizard prepare-config' on a scripts server while in a checkout
187 of this newest version. This will prepare a new version of the
188 configuration file based on the application's latest installer.
189 Manually merge back in any custom changes we may have made.
190 Check if any of the regular expressions need tweaking by inspecting
191 the configuration files for user-specific gunk, and modify
192 wizard.app.APPNAME accordingly.
194 12. Commit your changes, and tag as v1.2.3-scripts (or scripts2, if
195 you are amending an install without an upstream changes)
197 NOTE: These steps should be run on a scripts server
199 13. Test the new update procedure using our test scripts. See integration
200 tests for more information on how to do this.
202 http://scripts.mit.edu/wizard/testing.html#acceptance-tests
204 GET APPROVAL BEFORE PROCEEDING ANY FURTHER
206 NOTE: The following commands are to be run on not-backward.mit.edu.
207 You'll need to add daemon.scripts-security-upd to
208 scripts-security-upd to get bits to do this. Make sure you remove
209 these bits when you're done.
211 A 14. Run `wizard research appname`
212 which uses Git commands to check how many
213 working copies apply the change cleanly, and writes out a logfile
214 with the working copies that don't apply cleanly. It also tells
215 us about "corrupt" working copies, i.e. working copies that
216 have over a certain threshold of changes.
218 A 15. Run `wizard mass-upgrade appname`, which applies the update to all working
219 copies possible, and sends mail to users to whom the working copy
220 did not apply cleanly.
222 16. Run parallel-find.pl to update our inventory
224 * For mass importing into the repository, there are a few extra things:
226 * Many applications had patches associated with them. Be sure to
227 apply them, so later merges work better.
229 # the following operation might require -p1
230 patch -p0 < ../app-1.2.3/app-1.2.3.patch # [FIDDLY BIT]
232 * When running updates, if the patch has changed you will have to
233 do a special procedure for your merge:
235 git checkout pristine
236 # NOTE: Now, the tricky part (this is different from a real update)
237 git symbolic-ref HEAD refs/heads/master
238 # NOTE: Now, we think we're on the master branch, but we have
239 # pristine copy checked out
240 # NOTE: -p0 might need to be twiddled
241 patch -p0 < ../app-1.2.3/app-1.2.3.patch
243 # reconstitute .scripts directory
244 git checkout v1.2.2-scripts -- .scripts
246 # NOTE: Fake the merge
247 git rev-parse pristine > .git/MERGE_HEAD
249 You could also just try your luck with a manual merge using the patch
252 * The repository for a given application will contain the following files:
254 - The actual application's files, as from the official tarball
256 - A .scripts directory, which contains the following information:
258 * .scripts/.htaccess to prevent this directory from being accessed
261 * .scripts/old-version (optional) the old value of .scripts-version,
262 basically used for reverting an install to pre-migrated state.
264 * .scripts/lock (generated) which locks the autoinstall during an upgrade
266 * .scripts/version (deprecated)