blog/tags/bootsystem/index.html

   1 <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN"
   2         "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd">
   3 <html>
   4  <head>
   5   <title>Petter Reinholdtsen: Entries Tagged bootsystem</title>
   6   <link rel="stylesheet" type="text/css" media="screen" href="http://people.skolelinux.org/pere/blog/style.css">
   7   <link rel="alternate" title="RSS Feed" href="bootsystem.rss" type="application/rss+xml">
   8  </head>
   9  <body>
  10
  11  <div class="title">
  12   <h1>
  13        <a href="http://people.skolelinux.org/pere/blog/">Petter Reinholdtsen</a>
  14
  15   </h1>
  16
  17  </div>
  18
  19  <p>Entries tagged "bootsystem".</p>
  20
  21
  22
  23
  24 <div class="entry">
  25  <div class="title">
  26  <a href="http://people.skolelinux.org/pere/blog/Debian_boots_quicker_and_quicker.html">Debian boots quicker and quicker</a>
  27  </div>
  28  <div class="date">
  29   2009-06-24 21:40
  30  </div>
  31
  32  <div class="body">
  33
  34 <p>I spent Monday and tuesday this week in London with a lot of the
  35 people involved in the boot system on Debian and Ubuntu, to see if we
  36 could find more ways to speed up the boot system.  This was an Ubuntu
  37 funded
  38 <a href="https://wiki.ubuntu.com/FoundationsTeam/BootPerformance/DebianUbuntuSprint">developer
  39 gathering</a>. It was quite productive.  We also discussed the future
  40 of boot systems, and ways to handle the increasing number of boot
  41 issues introduced by the Linux kernel becoming more and more
  42 asynchronous and event base.  The Ubuntu approach using udev and
  43 upstart might be a good way forward.  Time will show.</p>
  44
  45 <p>Anyway, there are a few ways at the moment to speed up the boot
  46 process in Debian.  All of these should be applied to get a quick
  47 boot:</p>
  48
  49 <ul>
  50
  51 <li>Use dash as /bin/sh.</li>
  52
  53 <li>Disable the init.d/hwclock*.sh scripts and make sure the hardware
  54    clock is in UTC.</li>
  55
  56 <li>Install and activate the insserv package to enable
  57    <a href="http://wiki.debian.org/LSBInitScripts/DependencyBasedBoot">dependency
  58    based boot sequencing</a>, and enable concurrent booting.</li>
  59
  60 </ul>
  61
  62 These points are based on the Google summer of code work done by
  63 <a href="http://initscripts-ng.alioth.debian.org/soc2006-bootsystem/">Carlos
  64 Villegas</a>.
  65
  66 <p>Support for makefile-style concurrency during boot was uploaded to
  67 unstable yesterday.  When we tested it, we were able to cut 6 seconds
  68 from the boot sequence.  It depend on very correct dependency
  69 declaration in all init.d scripts, so I expect us to find edge cases
  70 where the dependences in some scripts are slightly wrong when we start
  71 using this.</p>
  72
  73 <p>On our IRC channel for this effort, #pkg-sysvinit, a new idea was
  74 introduced by Raphael Geissert today, one that could affect the
  75 startup speed as well.  Instead of starting some scripts concurrently
  76 from rcS.d/ and another set of scripts from rc2.d/, it would be
  77 possible to run a of them in the same process.  A quick way to test
  78 this would be to enable insserv and run 'mv /etc/rc2.d/S* /etc/rcS.d/;
  79 insserv'.  Will need to test if that work. :)</p>
  80
  81  </div>
  82  <div class="tags">
  83
  84
  85
  86   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>.
  87
  88  </div>
  89 </div>
  90 <div class="padding"></div>
  91
  92 <div class="entry">
  93  <div class="title">
  94  <a href="http://people.skolelinux.org/pere/blog/Taking_over_sysvinit_development.html">Taking over sysvinit development</a>
  95  </div>
  96  <div class="date">
  97   2009-07-22 23:00
  98  </div>
  99
 100  <div class="body">
 101
 102 <p>After several years of frustration with the lack of activity from
 103 the existing sysvinit upstream developer, I decided a few weeks ago to
 104 take over the package and become the new upstream.  The number of
 105 patches to track for the Debian package was becoming a burden, and the
 106 lack of synchronization between the distribution made it hard to keep
 107 the package up to date.</p>
 108
 109 <p>On the new sysvinit team is the SuSe maintainer Dr. Werner Fink,
 110 and my Debian co-maintainer Kel Modderman.  About 10 days ago, I made
 111 a new upstream tarball with version number 2.87dsf (for Debian, SuSe
 112 and Fedora), based on the patches currently in use in these
 113 distributions.  We Debian maintainers plan to move to this tarball as
 114 the new upstream as soon as we find time to do the merge.  Since the
 115 new tarball was created, we agreed with Werner at SuSe to make a new
 116 upstream project at <a href="http://savannah.nongnu.org/">Savannah</a>, and continue
 117 development there.  The project is registered and currently waiting
 118 for approval by the Savannah administrators, and as soon as it is
 119 approved, we will import the old versions from svn and continue
 120 working on the future release.</p>
 121
 122 <p>It is a bit ironic that this is done now, when some of the involved
 123 distributions are moving to upstart as a syvinit replacement.</p>
 124
 125  </div>
 126  <div class="tags">
 127
 128
 129
 130   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>, <a href="http://people.skolelinux.org/pere/blog/tags/nuug">nuug</a>.
 131
 132  </div>
 133 </div>
 134 <div class="padding"></div>
 135
 136 <div class="entry">
 137  <div class="title">
 138  <a href="http://people.skolelinux.org/pere/blog/Debian_has_switched_to_dependency_based_boot_sequencing.html">Debian has switched to dependency based boot sequencing</a>
 139  </div>
 140  <div class="date">
 141   2009-07-27 23:50
 142  </div>
 143
 144  <div class="body">
 145
 146 <p>Since this evening, with the upload of sysvinit version 2.87dsf-2,
 147 and the upload of insserv version 1.12.0-10 yesterday, Debian unstable
 148 have been migrated to using dependency based boot sequencing.  This
 149 conclude work me and others have been doing for the last three days.
 150 It feels great to see this finally part of the default Debian
 151 installation.  Now we just need to weed out the last few problems that
 152 are bound to show up, to get everything ready for Squeeze.</p>
 153
 154 <p>The next step is migrating /sbin/init from sysvinit to upstart, and
 155 fixing the more fundamental problem of handing the event based
 156 non-predictable kernel in the early boot.</p>
 157
 158  </div>
 159  <div class="tags">
 160
 161
 162
 163   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>, <a href="http://people.skolelinux.org/pere/blog/tags/nuug">nuug</a>.
 164
 165  </div>
 166 </div>
 167 <div class="padding"></div>
 168
 169 <div class="entry">
 170  <div class="title">
 171  <a href="http://people.skolelinux.org/pere/blog/Parallellizing_the_boot_in_Debian_Squeeze___ready_for_wider_testing.html">Parallellizing the boot in Debian Squeeze - ready for wider testing</a>
 172  </div>
 173  <div class="date">
 174   2010-05-06 23:25
 175  </div>
 176
 177  <div class="body">
 178
 179 <p>These days, the init.d script dependencies in Squeeze are quite
 180 complete, so complete that it is actually possible to run all the
 181 init.d scripts in parallell based on these dependencies.  If you want
 182 to test your Squeeze system, make sure
 183 <a href="http://wiki.debian.org/LSBInitScripts/DependencyBasedBoot">dependency
 184 based boot sequencing</a> is enabled, and add this line to
 185 /etc/default/rcS:</p>
 186
 187 <blockquote><pre>
 188 CONCURRENCY=makefile
 189 </pre></blockquote>
 190
 191 <p>That is it.  It will cause sysv-rc to use the startpar tool to run
 192 scripts in parallel using the dependency information stored in
 193 /etc/init.d/.depend.boot, /etc/init.d/.depend.start and
 194 /etc/init.d/.depend.stop to order the scripts.  Startpar is configured
 195 to try to start the kdm and gdm scripts as early as possible, and will
 196 start the facilities required by kdm or gdm as early as possible to
 197 make this happen.</p>
 198
 199 <p>Give it a try, and see if you like the result.  If some services
 200 fail to start properly, it is most likely because they have incomplete
 201 init.d script dependencies in their startup script (or some of their
 202 dependent scripts have incomplete dependencies).  Report bugs and get
 203 the package maintainers to fix it. :)</p>
 204
 205 <p>Running scripts in parallel could be the default in Debian when we
 206 manage to get the init.d script dependencies complete and correct.  I
 207 expect we will get there in Squeeze+1, if we get manage to test and
 208 fix the remaining issues.</p>
 209
 210 <p>If you report any problems with dependencies in init.d scripts to
 211 the BTS, please usertag the report to get it to show up at
 212 <a href="http://bugs.debian.org/cgi-bin/pkgreport.cgi?users=initscripts-ng-devel@lists.alioth.debian.org">the
 213 list of usertagged bugs related to this</a>.</p>
 214
 215  </div>
 216  <div class="tags">
 217
 218
 219
 220   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>.
 221
 222  </div>
 223 </div>
 224 <div class="padding"></div>
 225
 226 <div class="entry">
 227  <div class="title">
 228  <a href="http://people.skolelinux.org/pere/blog/systemd__an_interesting_alternative_to_upstart.html">systemd, an interesting alternative to upstart</a>
 229  </div>
 230  <div class="date">
 231   2010-05-13 22:20
 232  </div>
 233
 234  <div class="body">
 235
 236 <p>The last few days a new boot system called
 237 <a href="http://www.freedesktop.org/wiki/Software/systemd">systemd</a>
 238 has been
 239 <a href="http://0pointer.de/blog/projects/systemd.html">introduced</a>
 240
 241 to the free software world.  I have not yet had time to play around
 242 with it, but it seem to be a very interesting alternative to
 243 <a href="http://upstart.ubuntu.com/">upstart</a>, and might prove to be
 244 a good alternative for Debian when we are able to switch to an event
 245 based boot system.  Tollef is
 246 <a href="http://bugs.debian.org/580814">in the process</a> of getting
 247 systemd into Debian, and I look forward to seeing how well it work.  I
 248 like the fact that systemd handles init.d scripts with dependency
 249 information natively, allowing them to run in parallel where upstart
 250 at the moment do not.</p>
 251
 252 <p>Unfortunately do systemd have the same problem as upstart regarding
 253 platform support.  It only work on recent Linux kernels, and also need
 254 some new kernel features enabled to function properly.  This means
 255 kFreeBSD and Hurd ports of Debian will need a port or a different boot
 256 system.  Not sure how that will be handled if systemd proves to be the
 257 way forward.</p>
 258
 259 <p>In the mean time, based on the
 260 <a href="http://lists.debian.org/debian-devel/2010/05/msg00122.html">input
 261 on debian-devel@</a> regarding parallel booting in Debian, I have
 262 decided to enable full parallel booting as the default in Debian as
 263 soon as possible (probably this weekend or early next week), to see if
 264 there are any remaining serious bugs in the init.d dependencies.  A
 265 new version of the sysvinit package implementing this change is
 266 already in experimental.  If all go well, Squeeze will be released
 267 with parallel booting enabled by default.</p>
 268
 269  </div>
 270  <div class="tags">
 271
 272
 273
 274   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>, <a href="http://people.skolelinux.org/pere/blog/tags/nuug">nuug</a>.
 275
 276  </div>
 277 </div>
 278 <div class="padding"></div>
 279
 280 <div class="entry">
 281  <div class="title">
 282  <a href="http://people.skolelinux.org/pere/blog/Parallellized_boot_is_now_the_default_in_Debian_unstable.html">Parallellized boot is now the default in Debian/unstable</a>
 283  </div>
 284  <div class="date">
 285   2010-05-14 22:40
 286  </div>
 287
 288  <div class="body">
 289
 290 <p>Since this evening, parallel booting is the default in
 291 Debian/unstable for machines using dependency based boot sequencing.
 292 Apparently the testing of concurrent booting has been wider than
 293 expected, if I am to believe the
 294 <a href="http://lists.debian.org/debian-devel/2010/05/msg00122.html">input
 295 on debian-devel@</a>, and I concluded a few days ago to move forward
 296 with the feature this weekend, to give us some time to detect any
 297 remaining problems before Squeeze is frozen.  If serious problems are
 298 detected, it is simple to change the default back to sequential boot.
 299 The upload of the new sysvinit package also activate a new upstream
 300 version.</p>
 301
 302 More information about
 303 <a href="http://wiki.debian.org/LSBInitScripts/DependencyBasedBoot">dependency
 304 based boot sequencing</a> is available from the Debian wiki.  It is
 305 currently possible to disable parallel booting when one run into
 306 problems caused by it, by adding this line to /etc/default/rcS:</p>
 307
 308 <blockquote><pre>
 309 CONCURRENCY=none
 310 </pre></blockquote>
 311
 312 <p>If you report any problems with dependencies in init.d scripts to
 313 the BTS, please usertag the report to get it to show up at
 314 <a href="http://bugs.debian.org/cgi-bin/pkgreport.cgi?users=initscripts-ng-devel@lists.alioth.debian.org">the
 315 list of usertagged bugs related to this</a>.</p>
 316
 317  </div>
 318  <div class="tags">
 319
 320
 321
 322   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian edu">debian edu</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>.
 323
 324  </div>
 325 </div>
 326 <div class="padding"></div>
 327
 328 <div class="entry">
 329  <div class="title">
 330  <a href="http://people.skolelinux.org/pere/blog/Parallellized_boot_seem_to_hold_up_well_in_Debian_testing.html">Parallellized boot seem to hold up well in Debian/testing</a>
 331  </div>
 332  <div class="date">
 333   2010-05-27 23:55
 334  </div>
 335
 336  <div class="body">
 337
 338 <p>A few days ago, parallel booting was enabled in Debian/testing.
 339 The feature seem to hold up pretty well, but three fairly serious
 340 issues are known and should be solved:
 341
 342 <p><ul>
 343
 344 <li>The wicd package seen to
 345 <a href="http://bugs.debian.org/508289">break NFS mounting</a> and
 346 <a href="http://bugs.debian.org/581586">network setup</a> when
 347 parallel booting is enabled.  No idea why, but the wicd maintainer
 348 seem to be on the case.</li>
 349
 350 <li>The nvidia X driver seem to
 351 <a href="http://bugs.debian.org/583312">have a race condition</a>
 352 triggered more easily when parallel booting is in effect.  The
 353 maintainer is on the case.</li>
 354
 355 <li>The sysv-rc package fail to properly enable dependency based boot
 356 sequencing (the shutdown is broken) when old file-rc users
 357 <a href="http://bugs.debian.org/575080">try to switch back</a> to
 358 sysv-rc.  One way to solve it would be for file-rc to create
 359 /etc/init.d/.legacy-bootordering, and another is to try to make
 360 sysv-rc more robust.  Will investigate some more and probably upload a
 361 workaround in sysv-rc to help those trying to move from file-rc to
 362 sysv-rc get a working shutdown.</li>
 363
 364 </ul></p>
 365
 366 <p>All in all not many surprising issues, and all of them seem
 367 solvable before Squeeze is released.  In addition to these there are
 368 some packages with bugs in their dependencies and run level settings,
 369 which I expect will be fixed in a reasonable time span.</p>
 370
 371 <p>If you report any problems with dependencies in init.d scripts to
 372 the BTS, please usertag the report to get it to show up at
 373 <a href="http://bugs.debian.org/cgi-bin/pkgreport.cgi?users=initscripts-ng-devel@lists.alioth.debian.org">the
 374 list of usertagged bugs related to this</a>.</p>
 375
 376 <p>Update: Correct bug number to file-rc issue.</p>
 377
 378  </div>
 379  <div class="tags">
 380
 381
 382
 383   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian edu">debian edu</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>.
 384
 385  </div>
 386 </div>
 387 <div class="padding"></div>
 388
 389 <div class="entry">
 390  <div class="title">
 391  <a href="http://people.skolelinux.org/pere/blog/KDM_fail_at_boot_with_NVidia_cards___and_no_one_try_to_fix_it_.html">KDM fail at boot with NVidia cards - and no one try to fix it?</a>
 392  </div>
 393  <div class="date">
 394   2010-06-01 17:05
 395  </div>
 396
 397  <div class="body">
 398
 399 <p>It is strange to watch how a bug in Debian causing KDM to fail to
 400 start at boot when an NVidia video card is used is handled.  The
 401 problem seem to be that the nvidia X.org driver uses a long time to
 402 initialize, and this duration is longer than kdm is configured to
 403 wait.</p>
 404
 405 <p>I came across two bugs related to this issue,
 406 <a href="http://bugs.debian.org/583312">#583312</a> initially filed
 407 against initscripts and passed on to nvidia-glx when it became obvious
 408 that the nvidia drivers were involved, and
 409 <a href="http://bugs.debian.org/524751">#524751</a> initially filed against
 410 kdm and passed on to src:nvidia-graphics-drivers for unknown reasons.</p>
 411
 412 <p>To me, it seem that no-one is interested in actually solving the
 413 problem nvidia video card owners experience and make sure the Debian
 414 distribution work out of the box for these users.  The nvidia driver
 415 maintainers expect kdm to be set up to wait longer, while kdm expect
 416 the nvidia driver maintainers to fix the driver to start faster, and
 417 while they wait for each other I guess the users end up switching to a
 418 distribution that work for them.  I have no idea what the solution is,
 419 but I am pretty sure that waiting for each other is not it.</p>
 420
 421 <p>I wonder why we end up handling bugs this way.</p>
 422
 423  </div>
 424  <div class="tags">
 425
 426
 427
 428   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian edu">debian edu</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>.
 429
 430  </div>
 431 </div>
 432 <div class="padding"></div>
 433
 434 <div class="entry">
 435  <div class="title">
 436  <a href="http://people.skolelinux.org/pere/blog/Upstart_or_sysvinit___as_init_d_scripts_see_it.html">Upstart or sysvinit - as init.d scripts see it</a>
 437  </div>
 438  <div class="date">
 439   2010-06-06 23:55
 440  </div>
 441
 442  <div class="body">
 443
 444 <p>If Debian is to migrate to upstart on Linux, I expect some init.d
 445 scripts to migrate (some of) their operations to upstart job while
 446 keeping the init.d for hurd and kfreebsd.  The packages with such
 447 needs will need a way to get their init.d scripts to behave
 448 differently when used with sysvinit and with upstart.  Because of
 449 this, I had a look at the environment variables set when a init.d
 450 script is running under upstart, and when it is not.</p>
 451
 452 <p>With upstart, I notice these environment variables are set when a
 453 script is started from rcS.d/ (ignoring some irrelevant ones like
 454 COLUMNS):</p>
 455
 456 <blockquote><pre>
 457 DEFAULT_RUNLEVEL=2
 458 previous=N
 459 PREVLEVEL=
 460 RUNLEVEL=
 461 runlevel=S
 462 UPSTART_EVENTS=startup
 463 UPSTART_INSTANCE=
 464 UPSTART_JOB=rc-sysinit
 465 </pre></blockquote>
 466
 467 <p>With sysvinit, these environment variables are set for the same
 468 script.</p>
 469
 470 <blockquote><pre>
 471 INIT_VERSION=sysvinit-2.88
 472 previous=N
 473 PREVLEVEL=N
 474 RUNLEVEL=S
 475 runlevel=S
 476 </pre></blockquote>
 477
 478 <p>The RUNLEVEL and PREVLEVEL environment variables passed on from
 479 sysvinit are not set by upstart.  Not sure if it is intentional or not
 480 to not be compatible with sysvinit in this regard.</p>
 481
 482 <p>For scripts needing to behave differently when upstart is used,
 483 looking for the UPSTART_JOB environment variable seem to be a good
 484 choice.</p>
 485
 486  </div>
 487  <div class="tags">
 488
 489
 490
 491   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>.
 492
 493  </div>
 494 </div>
 495 <div class="padding"></div>
 496
 497 <div class="entry">
 498  <div class="title">
 499  <a href="http://people.skolelinux.org/pere/blog/Automatic_upgrade_testing_from_Lenny_to_Squeeze.html">Automatic upgrade testing from Lenny to Squeeze</a>
 500  </div>
 501  <div class="date">
 502   2010-06-11 22:50
 503  </div>
 504
 505  <div class="body">
 506
 507 <p>The last few days I have done some upgrade testing in Debian, to
 508 see if the upgrade from Lenny to Squeeze will go smoothly.  A few bugs
 509 have been discovered and reported in the process
 510 (<a href="http://bugs.debian.org/585410">#585410</a> in nagios3-cgi,
 511 <a href="http://bugs.debian.org/584879">#584879</a> already fixed in
 512 enscript and <a href="http://bugs.debian.org/584861">#584861</a> in
 513 kdebase-workspace-data), and to get a more regular testing going on, I
 514 am working on a script to automate the test.</p>
 515
 516 <p>The idea is to create a Lenny chroot and use tasksel to install a
 517 Gnome or KDE desktop installation inside the chroot before upgrading
 518 it.  To ensure no services are started in the chroot, a policy-rc.d
 519 script is inserted.  To make sure tasksel believe it is to install a
 520 desktop on a laptop, the tasksel tests are replaced in the chroot
 521 (only acceptable because this is a throw-away chroot).</p>
 522
 523 <p>A naive upgrade from Lenny to Squeeze using aptitude dist-upgrade
 524 currently always fail because udev refuses to upgrade with the kernel
 525 in Lenny, so to avoid that problem the file /etc/udev/kernel-upgrade
 526 is created.  The bug report
 527 <a href="http://bugs.debian.org/566000">#566000</a> make me suspect
 528 this problem do not trigger in a chroot, but I touch the file anyway
 529 to make sure the upgrade go well.  Testing on virtual and real
 530 hardware have failed me because of udev so far, and creating this file
 531 do the trick in such settings anyway.  This is a
 532 <a href="http://www.linuxquestions.org/questions/debian-26/failed-dist-upgrade-due-to-udev-config_sysfs_deprecated-nonsense-804130/">known
 533 issue</a> and the current udev behaviour is intended by the udev
 534 maintainer because he lack the resources to rewrite udev to keep
 535 working with old kernels or something like that.  I really wish the
 536 udev upstream would keep udev backwards compatible, to avoid such
 537 upgrade problem, but given that they fail to do so, I guess
 538 documenting the way out of this mess is the best option we got for
 539 Debian Squeeze.</p>
 540
 541 <p>Anyway, back to the task at hand, testing upgrades.  This test
 542 script, which I call <tt>upgrade-test</tt> for now, is doing the
 543 trick:</p>
 544
 545 <blockquote><pre>
 546 #!/bin/sh
 547 set -ex
 548
 549 if [ "$1" ] ; then
 550     desktop=$1
 551 else
 552     desktop=gnome
 553 fi
 554
 555 from=lenny
 556 to=squeeze
 557
 558 exec &lt; /dev/null
 559 unset LANG
 560 mirror=http://ftp.skolelinux.org/debian
 561 tmpdir=chroot-$from-upgrade-$to-$desktop
 562 fuser -mv .
 563 debootstrap $from $tmpdir $mirror
 564 chroot $tmpdir aptitude update
 565 cat > $tmpdir/usr/sbin/policy-rc.d &lt;&lt;EOF
 566 #!/bin/sh
 567 exit 101
 568 EOF
 569 chmod a+rx $tmpdir/usr/sbin/policy-rc.d
 570 exit_cleanup() {
 571     umount $tmpdir/proc
 572 }
 573 mount -t proc proc $tmpdir/proc
 574 # Make sure proc is unmounted also on failure
 575 trap exit_cleanup EXIT INT
 576
 577 chroot $tmpdir aptitude -y install debconf-utils
 578
 579 # Make sure tasksel autoselection trigger.  It need the test scripts
 580 # to return the correct answers.
 581 echo tasksel tasksel/desktop multiselect $desktop | \
 582     chroot $tmpdir debconf-set-selections
 583
 584 # Include the desktop and laptop task
 585 for test in desktop laptop ; do
 586     echo > $tmpdir/usr/lib/tasksel/tests/$test &lt;&lt;EOF
 587 #!/bin/sh
 588 exit 2
 589 EOF
 590     chmod a+rx $tmpdir/usr/lib/tasksel/tests/$test
 591 done
 592
 593 DEBIAN_FRONTEND=noninteractive
 594 DEBIAN_PRIORITY=critical
 595 export DEBIAN_FRONTEND DEBIAN_PRIORITY
 596 chroot $tmpdir tasksel --new-install
 597
 598 echo deb $mirror $to main > $tmpdir/etc/apt/sources.list
 599 chroot $tmpdir aptitude update
 600 touch $tmpdir/etc/udev/kernel-upgrade
 601 chroot $tmpdir aptitude -y dist-upgrade
 602 fuser -mv
 603 </pre></blockquote>
 604
 605 <p>I suspect it would be useful to test upgrades with both apt-get and
 606 with aptitude, but I have not had time to look at how they behave
 607 differently so far.  I hope to get a cron job running to do the test
 608 regularly and post the result on the web.  The Gnome upgrade currently
 609 work, while the KDE upgrade fail because of the bug in
 610 kdebase-workspace-data</p>
 611
 612 <p>I am not quite sure what kind of extract from the huge upgrade logs
 613 (KDE 167 KiB, Gnome 516 KiB) it make sense to include in this blog
 614 post, so I will refrain from trying.  I can report that for Gnome,
 615 aptitude report 760 packages upgraded, 448 newly installed, 129 to
 616 remove and 1 not upgraded and 1024MB need to be downloaded while for
 617 KDE the same numbers are 702 packages upgraded, 507 newly installed,
 618 193 to remove and 0 not upgraded and 1117MB need to be downloaded</p>
 619
 620 <p>I am very happy to notice that the Gnome desktop + laptop upgrade
 621 is able to migrate to dependency based boot sequencing and parallel
 622 booting without a hitch.  Was unsure if there were still bugs with
 623 packages failing to clean up their obsolete init.d script during
 624 upgrades, and no such problem seem to affect the Gnome desktop+laptop
 625 packages.</p>
 626
 627  </div>
 628  <div class="tags">
 629
 630
 631
 632   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian edu">debian edu</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>.
 633
 634  </div>
 635 </div>
 636 <div class="padding"></div>
 637
 638 <div class="entry">
 639  <div class="title">
 640  <a href="http://people.skolelinux.org/pere/blog/What_should_start_from__etc_rcS_d__in_Debian____almost_nothing.html">What should start from /etc/rcS.d/ in Debian? - almost nothing</a>
 641  </div>
 642  <div class="date">
 643   2011-07-30 14:00
 644  </div>
 645
 646  <div class="body">
 647
 648 <p>In the Debian boot system, several packages include scripts that
 649 are started from /etc/rcS.d/.  In fact, there is a bite more of them
 650 than make sense, and this causes a few problems.  What kind of
 651 problems, you might ask.  There are at least two problems.  The first
 652 is that it is not possible to recover a machine after switching to
 653 runlevel 1.  One need to actually reboot to get the machine back to
 654 the expected state.  The other is that single user boot will sometimes
 655 run into problems because some of the subsystems are activated before
 656 the root login is presented, causing problems when trying to recover a
 657 machine from a problem in that subsystem.  A minor additional point is
 658 that moving more scripts out of rcS.d/ and into the other rc#.d/
 659 directories will increase the amount of scripts that can run in
 660 parallel during boot, and thus decrease the boot time.</p>
 661
 662 <p>So, which scripts should start from rcS.d/.  In short, only the
 663 scripts that _have_ to execute before the root login prompt is
 664 presented during a single user boot should go there.  Everything else
 665 should go into the numeric runlevels.  This means things like
 666 lm-sensors, fuse and x11-common should not run from rcS.d, but from
 667 the numeric runlevels.  Today in Debian, there are around 115 init.d
 668 scripts that are started from rcS.d/, and most of them should be moved
 669 out.  Do your package have one of them?  Please help us make single
 670 user and runlevel 1 better by moving it.</p>
 671
 672 <p>Scripts setting up the screen, keyboard, system partitions
 673 etc. should still be started from rcS.d/, but there is for example no
 674 need to have the network enabled before the single user login prompt
 675 is presented.</p>
 676
 677 <p>As always, things are not so easy to fix as they sound.  To keep
 678 Debian systems working while scripts migrate and during upgrades, the
 679 scripts need to be moved from rcS.d/ to rc2.d/ in reverse dependency
 680 order, ie the scripts that nothing in rcS.d/ depend on can be moved,
 681 and the next ones can only be moved when their dependencies have been
 682 moved first.  This migration must be done sequentially while we ensure
 683 that the package system upgrade packages in the right order to keep
 684 the system state correct.  This will require some coordination when it
 685 comes to network related packages, but most of the packages with
 686 scripts that should migrate do not have anything in rcS.d/ depending
 687 on them.  Some packages have already been updated, like the sudo
 688 package, while others are still left to do.  I wish I had time to work
 689 on this myself, but real live constrains make it unlikely that I will
 690 find time to push this forward.</p>
 691
 692  </div>
 693  <div class="tags">
 694
 695
 696
 697   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>.
 698
 699  </div>
 700 </div>
 701 <div class="padding"></div>
 702
 703 <div class="entry">
 704  <div class="title">
 705  <a href="http://people.skolelinux.org/pere/blog/How_is_booting_into_runlevel_1_different_from_single_user_boots_.html">How is booting into runlevel 1 different from single user boots?</a>
 706  </div>
 707  <div class="date">
 708   2011-08-04 12:40
 709  </div>
 710
 711  <div class="body">
 712
 713 <p>Wouter Verhelst have some
 714 <a href="http://grep.be/blog/en/retorts/pere_kubuntu_boot">interesting
 715 pcomments and opinions</a> on my blog post on
 716 <a href="http://people.skolelinux.org/pere/blog/What_should_start_from__etc_rcS_d__in_Debian____almost_nothing.html">the
 717 need to clean up /etc/rcS.d/ in Debian</a> and my blog post about
 718 <a href="http://people.skolelinux.org/pere/blog/What_is_missing_in_the_Debian_desktop__or_why_my_parents_use_Kubuntu.html">the
 719 default KDE desktop in Debian</a>.  I only have time to address one
 720 small piece of his comment now, and though it best to address the
 721 misunderstanding he bring forward:</p>
 722
 723 <p><blockquote>
 724 Currently, a system admin has four options: [...] boot to a
 725 single-user system (by adding 'single' to the kernel command line;
 726 this runs rcS and rc1 scripts)
 727 </blockquote></p>
 728
 729 <p>This make me believe Wouter believe booting into single user mode
 730 and booting into runlevel 1 is the same.  I am not surprised he
 731 believe this, because it would make sense and is a quite sensible
 732 thing to believe.  But because the boot in Debian is slightly broken,
 733 runlevel 1 do not work properly and it isn't the same as single user
 734 mode.  I'll try to explain what is actually happing, but it is a bit
 735 hard to explain.</p>
 736
 737 <p>Single user mode is defined like this in /etc/inittab:
 738 "<tt>~~:S:wait:/sbin/sulogin</tt>".  This means the only thing that is
 739 executed in single user mode in sulogin.  Single user mode is a boot
 740 state "between" the runlevels, and when booting into single user mode,
 741 only the scripts in /etc/rcS.d/ are executed before the init process
 742 enters the single user state.  When switching to runlevel 1, the state
 743 is in fact not ending in runlevel 1, but it passes through runlevel 1
 744 and end up in the single user mode (see /etc/rc1.d/S03single, which
 745 runs "init -t1 S" to switch to single user mode at the end of runlevel
 746 1.  It is confusing that the 'S' (single user) init mode is not the
 747 mode enabled by /etc/rcS.d/ (which is more like the initial boot
 748 mode).</p>
 749
 750 <p>This summary might make it clearer.  When booting for the first
 751 time into single user mode, the following commands are executed:
 752 "<tt>/etc/init.d/rc S; /sbin/sulogin</tt>".  When booting into
 753 runlevel 1, the following commands are executed: "<tt>/etc/init.d/rc
 754 S; /etc/init.d/rc 1; /sbin/sulogin</tt>".  A problem show up when
 755 trying to continue after visiting single user mode.  Not all services
 756 are started again as they should, causing the machine to end up in an
 757 unpredicatble state.  This is why Debian admins recommend rebooting
 758 after visiting single user mode.</p>
 759
 760 <p>A similar problem with runlevel 1 is caused by the amount of
 761 scripts executed from /etc/rcS.d/.  When switching from say runlevel 2
 762 to runlevel 1, the services started from /etc/rcS.d/ are not properly
 763 stopped when passing through the scripts in /etc/rc1.d/, and not
 764 started again when switching away from runlevel 1 to the runlevels
 765 2-5.  I believe the problem is best fixed by moving all the scripts
 766 out of /etc/rcS.d/ that are not <strong>required</strong> to get a
 767 functioning single user mode during boot.</p>
 768
 769 <p>I have spent several years investigating the Debian boot system,
 770 and discovered this problem a few years ago.  I suspect it originates
 771 from when sysvinit was introduced into Debian, a long time ago.</p>
 772
 773  </div>
 774  <div class="tags">
 775
 776
 777
 778   Tags: <a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem</a>, <a href="http://people.skolelinux.org/pere/blog/tags/debian">debian</a>, <a href="http://people.skolelinux.org/pere/blog/tags/english">english</a>.
 779
 780  </div>
 781 </div>
 782 <div class="padding"></div>
 783
 784  <p style="text-align: right;"><a href="bootsystem.rss"><img src="http://people.skolelinux.org/pere/blog/xml.gif" alt="RSS Feed" width="36" height="14"></a></p>
 785
 786
 787
 788
 789 <div id="sidebar">
 790
 791 <h2>Archive</h2>
 792 <ul>
 793
 794 <li>2011
 795 <ul>
 796
 797 <li><a href="http://people.skolelinux.org/pere/blog/archive/2011/01/">January (16)</a></li>
 798
 799 <li><a href="http://people.skolelinux.org/pere/blog/archive/2011/02/">February (6)</a></li>
 800
 801 <li><a href="http://people.skolelinux.org/pere/blog/archive/2011/03/">March (6)</a></li>
 802
 803 <li><a href="http://people.skolelinux.org/pere/blog/archive/2011/04/">April (7)</a></li>
 804
 805 <li><a href="http://people.skolelinux.org/pere/blog/archive/2011/05/">May (3)</a></li>
 806
 807 <li><a href="http://people.skolelinux.org/pere/blog/archive/2011/06/">June (2)</a></li>
 808
 809 <li><a href="http://people.skolelinux.org/pere/blog/archive/2011/07/">July (7)</a></li>
 810
 811 <li><a href="http://people.skolelinux.org/pere/blog/archive/2011/08/">August (3)</a></li>
 812
 813 </ul></li>
 814
 815 <li>2010
 816 <ul>
 817
 818 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/01/">January (2)</a></li>
 819
 820 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/02/">February (1)</a></li>
 821
 822 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/03/">March (3)</a></li>
 823
 824 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/04/">April (3)</a></li>
 825
 826 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/05/">May (9)</a></li>
 827
 828 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/06/">June (14)</a></li>
 829
 830 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/07/">July (12)</a></li>
 831
 832 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/08/">August (13)</a></li>
 833
 834 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/09/">September (7)</a></li>
 835
 836 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/10/">October (9)</a></li>
 837
 838 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/11/">November (13)</a></li>
 839
 840 <li><a href="http://people.skolelinux.org/pere/blog/archive/2010/12/">December (12)</a></li>
 841
 842 </ul></li>
 843
 844 <li>2009
 845 <ul>
 846
 847 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/01/">January (8)</a></li>
 848
 849 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/02/">February (8)</a></li>
 850
 851 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/03/">March (12)</a></li>
 852
 853 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/04/">April (10)</a></li>
 854
 855 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/05/">May (9)</a></li>
 856
 857 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/06/">June (3)</a></li>
 858
 859 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/07/">July (4)</a></li>
 860
 861 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/08/">August (3)</a></li>
 862
 863 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/09/">September (1)</a></li>
 864
 865 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/10/">October (2)</a></li>
 866
 867 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/11/">November (3)</a></li>
 868
 869 <li><a href="http://people.skolelinux.org/pere/blog/archive/2009/12/">December (3)</a></li>
 870
 871 </ul></li>
 872
 873 <li>2008
 874 <ul>
 875
 876 <li><a href="http://people.skolelinux.org/pere/blog/archive/2008/11/">November (5)</a></li>
 877
 878 <li><a href="http://people.skolelinux.org/pere/blog/archive/2008/12/">December (7)</a></li>
 879
 880 </ul></li>
 881
 882 </ul>
 883
 884
 885
 886 <h2>Tags</h2>
 887 <ul>
 888
 889  <li><a href="http://people.skolelinux.org/pere/blog/tags/3d-printer">3d-printer (13)</a></li>
 890
 891  <li><a href="http://people.skolelinux.org/pere/blog/tags/amiga">amiga (1)</a></li>
 892
 893  <li><a href="http://people.skolelinux.org/pere/blog/tags/aros">aros (1)</a></li>
 894
 895  <li><a href="http://people.skolelinux.org/pere/blog/tags/bitcoin">bitcoin (2)</a></li>
 896
 897  <li><a href="http://people.skolelinux.org/pere/blog/tags/bootsystem">bootsystem (12)</a></li>
 898
 899  <li><a href="http://people.skolelinux.org/pere/blog/tags/debian">debian (53)</a></li>
 900
 901  <li><a href="http://people.skolelinux.org/pere/blog/tags/debian edu">debian edu (64)</a></li>
 902
 903  <li><a href="http://people.skolelinux.org/pere/blog/tags/digistan">digistan (7)</a></li>
 904
 905  <li><a href="http://people.skolelinux.org/pere/blog/tags/english">english (95)</a></li>
 906
 907  <li><a href="http://people.skolelinux.org/pere/blog/tags/fiksgatami">fiksgatami (12)</a></li>
 908
 909  <li><a href="http://people.skolelinux.org/pere/blog/tags/fildeling">fildeling (11)</a></li>
 910
 911  <li><a href="http://people.skolelinux.org/pere/blog/tags/intervju">intervju (10)</a></li>
 912
 913  <li><a href="http://people.skolelinux.org/pere/blog/tags/kart">kart (15)</a></li>
 914
 915  <li><a href="http://people.skolelinux.org/pere/blog/tags/ldap">ldap (8)</a></li>
 916
 917  <li><a href="http://people.skolelinux.org/pere/blog/tags/lenker">lenker (4)</a></li>
 918
 919  <li><a href="http://people.skolelinux.org/pere/blog/tags/ltsp">ltsp (1)</a></li>
 920
 921  <li><a href="http://people.skolelinux.org/pere/blog/tags/multimedia">multimedia (13)</a></li>
 922
 923  <li><a href="http://people.skolelinux.org/pere/blog/tags/norsk">norsk (131)</a></li>
 924
 925  <li><a href="http://people.skolelinux.org/pere/blog/tags/nuug">nuug (118)</a></li>
 926
 927  <li><a href="http://people.skolelinux.org/pere/blog/tags/open311">open311 (2)</a></li>
 928
 929  <li><a href="http://people.skolelinux.org/pere/blog/tags/opphavsrett">opphavsrett (21)</a></li>
 930
 931  <li><a href="http://people.skolelinux.org/pere/blog/tags/personvern">personvern (42)</a></li>
 932
 933  <li><a href="http://people.skolelinux.org/pere/blog/tags/reprap">reprap (11)</a></li>
 934
 935  <li><a href="http://people.skolelinux.org/pere/blog/tags/rfid">rfid (2)</a></li>
 936
 937  <li><a href="http://people.skolelinux.org/pere/blog/tags/robot">robot (4)</a></li>
 938
 939  <li><a href="http://people.skolelinux.org/pere/blog/tags/rss">rss (1)</a></li>
 940
 941  <li><a href="http://people.skolelinux.org/pere/blog/tags/sikkerhet">sikkerhet (23)</a></li>
 942
 943  <li><a href="http://people.skolelinux.org/pere/blog/tags/sitesummary">sitesummary (3)</a></li>
 944
 945  <li><a href="http://people.skolelinux.org/pere/blog/tags/standard">standard (24)</a></li>
 946
 947  <li><a href="http://people.skolelinux.org/pere/blog/tags/stavekontroll">stavekontroll (1)</a></li>
 948
 949  <li><a href="http://people.skolelinux.org/pere/blog/tags/stortinget">stortinget (2)</a></li>
 950
 951  <li><a href="http://people.skolelinux.org/pere/blog/tags/surveillance">surveillance (9)</a></li>
 952
 953  <li><a href="http://people.skolelinux.org/pere/blog/tags/video">video (20)</a></li>
 954
 955  <li><a href="http://people.skolelinux.org/pere/blog/tags/vitenskap">vitenskap (1)</a></li>
 956
 957  <li><a href="http://people.skolelinux.org/pere/blog/tags/web">web (16)</a></li>
 958
 959 </ul>
 960
 961 </div>
 962 </body>
 963 </html>