5 - Keep my sanity when upgrading 1000 installs
6 - Custom merge algo: absolute php.ini symlinks to relative symlinks (this
7 does not seem to have been a problem in practice)
8 - Prune -7 call errors and automatically reprocess them (with a
9 strike out counter of 3)--this requires better error parsing.
10 - IOError should be aggregated, right now contains custom string
11 that makes this not possible. Partition on a colon.
12 - Replace gaierror with a more descriptive name (this is a DNS error)
13 - Stronger skips means that backup failures should also be avoided
14 - Distinguish between types of backup failures
15 - Ignore empty blacklists; they should all have reasons
16 - wizard upgrade should have different exit codes for merge failure
19 - Distinguish from logging and reporting (so we can easily send mail
21 - Figure out a way of collecting blacklist data from .scripts/blacklisted
22 and aggregate it together
24 - Let users use Wizard when ssh'ed into Scripts
25 - Make single user mass-migrate work when not logged in as root
27 - Make the rest of the world use Wizard
28 - Make parallel-find.pl use `sudo -u username git describe --tags`
29 to determine version. Make parallel-find.pl have this have greater
30 precedence. This also means, however, that we get
31 full mediawiki-1.2.3-2-abcdef names (Have patch, pending testing and commit)
32 - Make deployed installer use 'wizard install' /or/ do a migration
33 after doing a normal install (the latter makes it easier
36 - Pre-emptively check if daemon/scripts-security-upd
37 is not on scripts-security-upd list (/mit/moira/bin/blanche)
39 - Redo Wordpress conversion, with an eye for automating everything
40 possible (such as downloading the tarball and unpacking)
43 - Genericize callAsUser and drop_priviledges in shell
44 - Summary script should be more machine friendly, and should not
45 output summary charts when I increase specificity
46 - Summary script should do something intelligent when distinguishing
47 between old-style and new-style installs
48 - Report code in wizard/command/__init__.py is ugly as sin. Also,
49 the Report object should operate at a higher level of abstraction
50 so we don't have to manually increment fails. (in fact, that should
51 probably be called something different). The by-percent errors should
53 - Indents in upgrade.py are getting pretty ridiculous; more breaking
54 into functions is probably a good idea
55 - Move resolutions in mediawiki.py to a text file? (the parsing overhead
57 - Investigate QuotaParseErrors
58 - If a process is C-ced, it can result in a upgrade that has
59 an updated filesystem but not updated database. Make this more
61 - PHP end of file allows omitted semicolon, can result in parse error
62 if merge resolutions aren't careful.
65 - Don't use the scripts heuristics unless we're on scripts with the
66 AFS patch. Check with `fs sysname`
67 - Make 'wizard summary' generate nice pretty graphs of installs by date
68 (more histograms, will need to check actual .scripts-version files.)
69 - It should be able to handle installs like Django where there's a component
70 that gets installed in web_scripts and another directory that gets installed
72 - ACLs is a starting point for sending mail to users, but it has
73 several failure modes:
74 - Old maintainers who don't care who are still on the ACL
75 - Private AFS groups that aren't mailing lists and that we
77 A question is whether or not sending mail actually helps us:
78 many users will probably have to come back to us for help; many
79 other users won't care.
81 PULLING OUT CONFIGURATION FILES IN AN AUTOMATED MANNER
83 advancedpoll: Template file to fill out
84 django: Noodles of template files
85 gallery2: Multistage install process
87 mediawiki: One-step install process
88 phpbb: Multistage install process
89 phpical: Template file
92 wordpress: Multistage install process
94 PHILOSOPHY ABOUT LOGGING
96 Logging is most useful when performing a mass run. This
97 includes things such as mass-migration as well as when running
98 summary reports. An interesting property about mass-migration
99 or mass-upgrade, however, is that if they fail, they are
100 idempotent, so an individual case can be debugged simply running
101 the single-install equivalent with --debug on. (This, indeed,
102 may be easier to do than sifting through a logfile).
104 It is a different story when you are running a summary report:
105 you are primarily bound by your AFS cache and how quickly you can
106 iterate through all of the autoinstalls. Checking if a file
107 exists on a cold AFS cache may
108 take several minutes to perform; on a hot cache the same report
109 may take a mere 3 seconds. When you get to more computationally
110 expensive calculations, however, even having a hot AFS cache
111 is not enough to cut down your runtime.
113 There are certain calculations that someone may want to be
114 able to perform on manipulated data. As such, this data should
115 be cached on disk, if the process for extracting this data takes
116 a long time. Also, for usability sake, Wizard should generate
117 the common case reports.
119 Ensuring that machine parseable reports are made, and then making
120 the machinery to reframe this data, increases complexity. Therefore,
121 the recommendation is to assume that if you need to run iteratively,
122 you'll have a hot AFS cache at your fingerprints, and if that's not
123 fast enough, then cache the data.
125 COMMIT MESSAGE FIELDS:
127 Installed-by: username@hostname
128 Pre-commit-by: Real Name <username@mit.edu>
129 Upgraded-by: Real Name <username@mit.edu>
130 Migrated-by: Real Name <username@mit.edu>
131 Wizard-revision: abcdef1234567890
132 Wizard-args: /wizard/bin/wizard foo bar baz
136 Committer: Real Name <username@mit.edu>
137 Author: lockername locker <lockername@scripts.mit.edu>
141 - It is not expected or required for update scripts to exist for all
142 intervening versions that were present pre-migration; only for it
143 to work on the most recent migration.
145 - Currently all repositories are initialized with --shared, which
146 means they have basically ~no space footprint. However, it
147 also means that /mit/scripts/wizard/srv MUST NOT lose revs after
150 - Full fledged logging options. Namely:
151 x all loggers (delay implementing this until we actually have debug stmts)
153 - debug => loglevel = DEBUG
155 - default is WARNING (see below for exception)
156 - verbose => loglevel = INFO
157 x file logger (creates a dir and lots of little logfiles)
159 - log-file => loglevel = INFO
163 * Some parts of the infrastructure will not be touched, although I plan
164 on documenting them. Specifically, we will be keeping:
166 - parallel-find.pl, and the resulting
167 /mit/scripts/.htaccess/scripts/sec-tools/store/scriptslist
169 * The new procedure for generating an update is as follows:
170 (check out the mass-migration instructions for something in this spirit,
171 although uglier in some ways; A indicates the step /should/ be automated)
173 0. ssh into not-backward, temporarily give the daemon.scripts-security-upd
174 bits by blanching it on system:scripts-security-upd, and run parallel-find.pl
176 1. Have the Git repository and working copy for the project on hand.
178 /- wizard prepare-pristine --
180 A 2. Checkout the pristine branch
182 A 3. Remove all files from the working copy. Use `wipe-working-dir`
184 A 4. Download the new tarball
186 A 5. Extract the tarball over the working copy (`cp -R a/. b` works well,
187 remember that the working copy is empty; this needs some intelligent
190 A 6. Check for empty directories and add stub files as necessary.
191 Use `preserve-empty-dir`
195 7. Git add it all, and then commit as a new pristine version (v1.2.3)
197 8. Checkout the master branch
199 9. [FOR EXISTING REPOSITORIES]
200 Merge the pristine branch in. Resolve any conflicts that our
201 patches have with new changes. Do NOT let Git auto-commit it
202 with --no-commit (otherwise, you want to git commit --amend
203 to keep our history clean
205 [FOR NEW REPOSITORIES]
206 Check if any patches are needed to make the application work
207 on Scripts (ideally, it shouldn't.
209 /- wizard prepare-new --
211 Currently not used for anything besides parallel-find.pl, but
212 we reserve the right to place files in here in the future.
215 A echo "Deny from all" > .scripts/.htaccess
219 10. Check if there are any special update procedures, and update
220 the wizard.app.APPNAME module accordingly (or create it, if
223 11. Run 'wizard prepare-config' on a scripts server while in a checkout
224 of this newest version. This will prepare a new version of the
225 configuration file based on the application's latest installer.
226 Manually merge back in any custom changes we may have made.
227 Check if any of the regular expressions need tweaking by inspecting
228 the configuration files for user-specific gunk, and modify
229 wizard.app.APPNAME accordingly.
231 12. Commit your changes, and tag as v1.2.3-scripts (or scripts2, if
232 you are amending an install without an upstream changes)
234 NOTE: These steps should be run on a scripts server
236 13. Test the new update procedure using our test scripts. See integration
237 tests for more information on how to do this.
239 http://scripts.mit.edu/wizard/testing.html#acceptance-tests
241 GET APPROVAL BEFORE PROCEEDING ANY FURTHER
243 NOTE: The following commands are to be run on not-backward.mit.edu.
244 You'll need to add daemon.scripts-security-upd to
245 scripts-security-upd to get bits to do this. Make sure you remove
246 these bits when you're done.
248 A 14. Run `wizard research appname`
249 which uses Git commands to check how many
250 working copies apply the change cleanly, and writes out a logfile
251 with the working copies that don't apply cleanly. It also tells
252 us about "corrupt" working copies, i.e. working copies that
253 have over a certain threshold of changes.
255 A 15. Run `wizard mass-upgrade appname`, which applies the update to all working
256 copies possible, and sends mail to users to whom the working copy
257 did not apply cleanly.
259 16. Run parallel-find.pl to update our inventory
261 * For mass importing into the repository, there are a few extra things:
263 * Many applications had patches associated with them. Be sure to
264 apply them, so later merges work better.
266 # the following operation might require -p1
267 patch -p0 < ../app-1.2.3/app-1.2.3.patch # [FIDDLY BIT]
269 * When running updates, if the patch has changed you will have to
270 do a special procedure for your merge:
272 git checkout pristine
273 # NOTE: Now, the tricky part (this is different from a real update)
274 git symbolic-ref HEAD refs/heads/master
275 # NOTE: Now, we think we're on the master branch, but we have
276 # pristine copy checked out
277 # NOTE: -p0 might need to be twiddled
278 patch -p0 < ../app-1.2.3/app-1.2.3.patch
280 # reconstitute .scripts directory
281 git checkout v1.2.2-scripts -- .scripts
283 # NOTE: Fake the merge
284 git rev-parse pristine > .git/MERGE_HEAD
286 You could also just try your luck with a manual merge using the patch
289 * The repository for a given application will contain the following files:
291 - The actual application's files, as from the official tarball
293 - A .scripts directory, with the intent of holding Scripts specific files
294 if they become necessary.
296 * .scripts/lock (generated) which locks an autoinstall during upgrade