1 2009-04-10 Andreas J. Koenig <andk@cpan.org>
3 * Todo: rename rrr-dirtyupdate to rrr-update and add an option --dirty
4 so we can use it for normal update with a single file.
6 * Bug? When the dirtymark changes then the second tier hosts quickly
7 reset their DONE state and restart mirroring. But then they mirror
8 potentially outdated index files with inconsistent dirtymark because the
9 first tier box may be a bit behind on the large recentfiles.
11 This means we need a brake when iterating over the recentfiles that
12 refuses to use a recentfile with the wrong dirtymark.
14 2009-04-08 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
16 * install the bin/* files and talk about rrr-overview somewhere.
18 * Ask B. Hansen asks for tempdir accessor such that tempfiles get
19 created outside the target tree.
21 * Barbie asks for better logging capabilities
23 2009-03-29 Andreas J. Koenig <andk@cpan.org>
25 * known bug: cannot configure the Recentfile objects to keep delete
26 events. How to call the accessor for that? keep_delete_objects_forever?
28 Sounds acceptable. FIXED (but untested).
30 * known bug: keeps temporary index files lying around.
32 2009-03-24 Andreas J. Koenig <andk@cpan.org>
36 for ( @{$rrr->recentfiles} ) { $_->verbose(0) }
40 2009-03-22 Andreas J. Koenig <andk@cpan.org>
42 * k81 is again client of k75 with different parameters.
44 * broken now: (1) seeding and unseeding: rmirror is seeding and talking
45 about it all the time and nobody reacts accordingly; (2) lots of temp
46 files get created and not removed; culprit the new call to
47 get_remote_recentfile_as_tempfile within Recent.pm; the manpage says the
48 caller has to remove the tempfile after use.
50 Need the drawing board.
52 PARTLY FIXED: retracting the idea to call
53 get_remote_recentfile_as_tempfile from Recent.pm and moving around seed
54 and unseed such that uptodate now means "we have mirrored this
55 recentfile's bunch of files" and seeded means "somebody believes the
56 index file for this rf needs be refreshed". This seems to work now.
58 Still collecting temp files that nobody cares to depollute.
60 2009-03-21 Andreas J. Koenig <andk@cpan.org>
62 * Bug?: should it be harder than it is atm to set the timestamp to the
65 * bug with native integers:
69 path: id/D/DE/DELTA/Crypt-Rijndael_PP-0.03.readme
73 path: id/L/LG/LGODDARD/Tk-Wizard-2.124.readme
76 Native integer broke when native math was turned off. FIXED
78 * Bug: something between id/P/PH/PHISH/CGI-XMLApplication-1.1.2.readme
79 and id/C/CH/CHOGAN/HTML-WWWTheme-1.06.readme. Mirroring the Z file loops
82 Yes, records out of order:
85 447004 epoch: 995885533
86 447005 path: id/P/PH/PHISH/CGI-XMLApplication_0.9.3.readme
89 447008 epoch: 995890358
90 447009 path: id/H/HD/HDIAS/Mail-Cclient-1.3.readme
93 447012 epoch: 995892221
94 447013 path: id/H/HD/HDIAS/Mail-Cclient-1.3.tar.gz
97 FIXED with sanity check and later with the integer fix.
99 * Bug: want the index files in a .recent directory
101 * Bug: lots of dot files are not deleted in time
103 * possible test case: can a delete change the timestamp? This would
104 probably break the order of events.
106 2009-03-20 Andreas J. Koenig <andk@cpan.org>
108 * 1233701831.34486 what's so special about this number/string? It
110 id/G/GR/GRODITI/MooseX-Emulate-Class-Accessor-Fast-0.00800.tar.gz and
111 atm lives in Y,Q, and Z.
113 It is the first entry after id/--skip-locking which has timestamp
114 1234164228.11325 which represents a file that doesn't exist anymore.
118 Sync 1237531537 (31547/33111/Z) id/J/JH/JHI/String-Approx-2.7.tar.gz ...
119 _bigfloatcmp called with l[1237505213.21133]r[UNDEF]: but both must be defined at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 76
120 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 131
121 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatlt(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 110
122 File::Rsync::Mirror::Recentfile::Done::covered('File::Rsync::Mirror::Recentfile::Done=HASH(0x8857fb4)', 1237505213.21133, 0.123456789) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile.pm line 2041
123 File::Rsync::Mirror::Recentfile::uptodate('File::Rsync::Mirror::Recentfile=HASH(0x8533a2c)') called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recent.pm line 536
124 File::Rsync::Mirror::Recent::rmirror('File::Rsync::Mirror::Recent=HASH(0x82ef3d0)', 'skip-deletes', 1) called at /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl line 27
125 at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 76
126 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 131
127 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatlt(1237505213.21133, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 110
128 File::Rsync::Mirror::Recentfile::Done::covered('File::Rsync::Mirror::Recentfile::Done=HASH(0x8857fb4)', 1237505213.21133, 0.123456789) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile.pm line 2041
129 File::Rsync::Mirror::Recentfile::uptodate('File::Rsync::Mirror::Recentfile=HASH(0x8533a2c)') called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recent.pm line 536
130 File::Rsync::Mirror::Recent::rmirror('File::Rsync::Mirror::Recent=HASH(0x82ef3d0)', 'skip-deletes', 1) called at /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl line 27
132 FIXED, it was the "--skip-locking" file where manual intervention was participating
134 * bug on the mirroring slave: when the dirtymark gets increased we
135 probably do not reset the done intervals. The mirrorer stays within
136 tight bounds where it tries to sync with upstream and never seems to
137 finish. In the debugging state file I see lots of identical intervals
138 that do not get collapsed. When I restart the mirrorer it dies with:
140 Sync 1237507989 (227/33111/Z) id/X/XI/XINMING/Catalyst-Plugin-Compress.tar.gz ...
141 _bigfloatcmp called with l[1237400817.94363]r[UNDEF]: but both must be defined at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 76
142 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatcmp(1237400817.94363, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/FakeBigFloat.pm line 101
143 File::Rsync::Mirror::Recentfile::FakeBigFloat::_bigfloatge(1237400817.94363, undef) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 226
144 File::Rsync::Mirror::Recentfile::Done::_register_one('File::Rsync::Mirror::Recentfile::Done=HASH(0x84c6af8)', 'HASH(0xb693f2dc)') called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile/Done.pm line 200
145 File::Rsync::Mirror::Recentfile::Done::register('File::Rsync::Mirror::Recentfile::Done=HASH(0x84c6af8)', 'ARRAY(0x8c54bfc)', 'ARRAY(0xb67e618c)') called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile.pm line 1044
146 File::Rsync::Mirror::Recentfile::_mirror_item('File::Rsync::Mirror::Recentfile=HASH(0x84abcb0)', 227, 'ARRAY(0x8c54bfc)', 33110, 'File::Rsync::Mirror::Recentfile::Done=HASH(0x84c6af8)', 'HASH(0x84abd8c)', 'ARRAY(0x839cb64)', 'HASH(0x839c95c)', 'HASH(0xb6a30f0c)', ...) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recentfile.pm line 992
147 File::Rsync::Mirror::Recentfile::mirror('File::Rsync::Mirror::Recentfile=HASH(0x84abcb0)', 'piecemeal', 1, 'skip-deletes', 1) called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recent.pm line 564
148 File::Rsync::Mirror::Recent::_rmirror_mirror('File::Rsync::Mirror::Recent=HASH(0x84ab6d4)', 7, 'HASH(0x8499488)') called at /home/k/sources/rersyncrecent/lib/File/Rsync/Mirror/Recent.pm line 532
149 File::Rsync::Mirror::Recent::rmirror('File::Rsync::Mirror::Recent=HASH(0x84ab6d4)', 'skip-deletes', 1) called at /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl line 27
151 and debugging stands at
157 epoch: 1237400802.5789
158 path: id/J/JO/JOHND/CHECKSUMS
161 epoch: 1237400807.97514
162 path: id/M/MI/MIYAGAWA/CHECKSUMS
165 epoch: 1237400817.94363
166 path: id/X/XI/XINMING/Catalyst-Plugin-Compress.tar.gz
174 and it is reproducable.
176 Why does the mirrorer not fetch a newer Z file? It is 18 hours old while
177 pause has a fresh one.
179 FIXED, it was the third anded term in each of the ifs in the IV block in
180 _register_one: with that we make sure that we do not stamp on valuable
183 2009-03-17 Andreas J. Koenig <andk@cpan.org>
185 * done: verified the existence of the floating point bug in bleadperl
186 and verified that switching from YAML::Syck to YAML::XS does not resolve
189 BTW, the switch was doable with
191 perl -i~ -pe 's/Syck/XS/g' lib/**/*.pm t/*.t
193 and should be considered as a separate TODO
195 * todo: integrate a dirty update with two aggregate calls before
196 unlocking for frictionless dirtying
198 * todo: start the second rsync daemon on pause
200 * todo: move index files to .recent: this cannot simply be done by
201 setting filenameroot to .recent/RECENT. Other parts of the modules rely
202 on the fact that dirname(recentfile) is the root of the mirrored tree.
204 2009-03-16 Andreas J. Koenig <andk@cpan.org>
206 * What was the resolution of the mirror.pl delete hook bug? Do we call
207 the delete hook when pause removes a file from MUIR?
209 * Today on pause: Updating 2a13fba..29f284d and installing it for
210 /usr/local/perl-5.10.0{,-RC2}
212 TURUGINA/Set-Intersection-0.01.tar.gz was the last upload before this
213 action and G/GW/GWILLIAMS/RDF-Query-2.100_01.tar.gz the first after it
215 2009-03-15 Andreas J. Koenig <andk@cpan.org>
217 * currently recent_events has the side effect of setting dirtymark
218 because it forces a read on the file. That should be transparent, so
219 that the dirtymark call always forces a cache-able(?) read.
221 * The bug below is -- after a lot of trying -- not reproducible on a
222 small script, only in the large test script. The closest to the output
228 my $x = "01237123229.8814";
233 ($l,$r) = ($1,$2) if $x =~ /(.)(.+)/;
235 $l = "1237123231.22458";
236 $r = "1237123231.22458";
238 Devel::Peek::Dump $l;
239 Devel::Peek::Dump $r;
240 Devel::Peek::Dump $x = $l <=> $r;
244 The checked in state at c404a85 fails the test with my
245 /usr/local/perl-5.10-uld/bin/perl on 64bit but curiously not with
246 /usr/local/perl-5.10-g/bin/perl. So it seems the behaviour is not even
247 in the test script always consistent.
249 * Todo: write a test that inserts a second dirty file with an already
250 existing timestamp. DONE
252 * Bug in perl 5.10 on my 64bit box:
254 DB<98> Devel::Peek::Dump $l
255 SV = PVMG(0x19e0450) at 0x142a550
257 FLAGS = (PADMY,NOK,POK,pNOK,pPOK)
259 NV = 1237123231.22458
260 PV = 0x194ce70 "1237123231.22458"\0
264 DB<99> Devel::Peek::Dump $r
265 SV = PVMG(0x19e0240) at 0x142a3e8
267 FLAGS = (PADMY,POK,pPOK)
270 PV = 0x19ff900 "1237123231.22458"\0
274 DB<100> Devel::Peek::Dump $l <=> $r
275 SV = IV(0x19ea6e8) at 0x19ea6f0
277 FLAGS = (PADTMP,IOK,pIOK)
280 DB<101> Devel::Peek::Dump $l
281 SV = PVMG(0x19e0450) at 0x142a550
283 FLAGS = (PADMY,NOK,POK,pNOK,pPOK)
285 NV = 1237123231.22458
286 PV = 0x194ce70 "1237123231.22458"\0
290 DB<102> Devel::Peek::Dump $r
291 SV = PVMG(0x19e0240) at 0x142a3e8
293 FLAGS = (PADMY,NOK,POK,pIOK,pNOK,pPOK)
295 NV = 1237123231.22458
296 PV = 0x19ff900 "1237123231.22458"\0
300 Retry with uselongdouble gives same effect. Not reproducable on 32bit box (k75).
302 * Todo: reset "done" or "covered" and "minmax" after a dirty operation?
304 2009-03-11 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
306 * $obj->merge ($other) needs to learn about equal epoch which may happen
307 since dirty_epoch intruded.
309 * Wontfix anytime soon: I think we currently do not support mkdir. Only
312 2009-01-01 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
314 * Todo: continue working on update(...,$dirty_epoch). It must be
315 followed by a fast_aggregate!
317 2008-12-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
319 * maybe we need a closest_entry or fitting_interval or something like
320 that. We want to merge an event into the middle of some recentfile.
321 First we do not know which file, then we do not know where to lock,
322 where to enter the new item, when and where to correct the dirtymark.
324 So my thought is we should first find which file.
326 Another part of my brain answers: what would happen if we would enter
327 the new file into the smallest file just like an ordinary new event,
328 just as an old event?
330 (1) we would write a duplicate timestamp? No, this would be easy to
333 (2) we would make the file large quickly? Yes, but so what? We are
334 changing the dirtymark, so are willing to disturb the downstream hosts.
336 2008-11-22 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
338 * 10705 root 17 0 725m 710m 1712 S 0.0 46.8 834:56.05 /home/src/perl/repoperls/installed-perls/perl/pVNtS9N/perl-5.8.0@32642/bin/perl -Ilib /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl
342 https://rt.cpan.org/Ticket/Display.html?id=41199
344 * bzcat uploads.csv.bz2 | perl -F, -nale '$Seen{$F[-1]}++ and print'
346 Strangest output being HAKANARDO who managed to upload
348 Here is a better oneliner that includes also the first line of each
351 bzcat uploads.csv.bz2 | perl -MYAML::Syck -F, -nale '$F[-1]=~s/\s+\z//; push @{$Seen{$F[-1]}||=[]},$_; END {for my $k (keys %Seen){ delete $Seen{$k} if @{$Seen{$k}}==1; } print YAML::Syck::Dump(\%Seen)}'
353 2008-10-31 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
355 * memory leak in the syncher? It currently weighs 100M.
359 root 10705 1.0 4.9 80192 76596 pts/32 S+ Nov02 24:05 /home/src/perl/repoperls/installed-perls/perl/pVNtS9N/perl-5.8.0@32642/bin/perl -Ilib /home/k/sources/CPAN/GIT/trunk/bin/testing-rmirror.pl
362 2008-10-29 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
364 * lookup by epoch and by path and use this ability on the pause to never
365 again register a file twice that doesn't need it. Let's call it
368 * after the dirtymark is done: fill up recentfiles with fake (historic)
369 entries; fill up with individual corrections; algorithm maybe to be done
370 with bigfloat so that we can always place something in the middle
371 between two entries. Before we must switch to bigfloat we could try to
372 use Data::Float::nextup to get the.
374 * Inotify2 on an arbitrary tree and then play with that instead of PAUSE
377 * dirtymark now lives in Recentfile, needs to be used in rmirror.
379 * find out why the downloader died after a couple of hours without a net
380 connection. Write a test that survives the not-existence of the other
383 2008-10-15 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
385 * reconsider the HTTP epoch only. Not the whole thing over HTTP because
386 it makes less sense with tight coupling for secondary files. But asking
387 the server what the current epoch is might be cheaper on HTTP than on
388 rsync. (Needs to be evaluated)
390 * remove the 0.00 from the verbose overview in the Merged column in the
393 * write tests that expose the problems of the last few days: cascading
394 client/server roles, tight coupling for secondary RFs, deletes after
397 * Some day we might want to have policy options for the slave:
398 tight/loose/no coupling with upstream for secondary RFs. tight is what
399 we have now. loose would wait until a gap occurs that can be closed.
401 2008-10-14 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
403 * revisit all $rfs->[$i+1] places if they now make sense still
405 2008-10-11 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
407 * another bug is the fact that the mirror command deletes files before
408 it unhides the index file, thus confusing downstream slaves. We must not
409 delete before unhiding and must delete after unhiding. FIXED.
411 * new complication about the slave that is playing a server role.
412 Currently we mirror from newest to oldest with a hidden temporary file
413 as index. And when one file is finished, we unhide the index file.
414 Imagine the cascading server/slave is dead for a day. It then starts
415 mirroring again with the freshest thing and unhides the freshest index
416 file when it has worked through it. In that moment it exposes a time
417 hole. Because it now works on the second recentfile which is still
420 We currently do nothing special to converge after such a drop out. At
421 least not intentionally and robustly and thought through.
423 The algorithm we use to seed the next file needs quite a lot of more
424 robustness than it currently has. Something to do with looking at the
425 merged element of the next rf and when it has dropped off, we seed
426 immediately. And if it ramains dropped off, we seed again, of course.
428 Nope, looking from smaller to larger RFS we look at the merged element
429 of this RF and at the minmax/max element of the next RF. If that
430 $rf[next]->{minmax}{max} >= $rf[this]->{merged}{epoch}, then we can stop
433 And we need a public accessor seed and unseed or seeded. But not the mix
434 of public and private stuff that then is used behind the back.
436 And then the secondary* stuff must go.
438 And we must understand what the impact is on the DONE system. Can it go
439 unnoticed that there was a hole? And could the DONE system have decided
440 the hole is covered? This should be testable with three directories where
441 the middle stops working for a while. Done->merge is suspicious, we must
442 stop it from merging non-conflatable neighbors due to broken continuity.
446 2008-10-10 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
448 * Slaven suggests to have the current epoch or the whole current
449 recentfile available from the HTTP server and take it away with
450 keepalive. This direction goes the granularity down to subseconds.
452 We might want to rewrite everything to factor out transport and allow
453 the whole thing to run via HTTP.
455 2008-10-09 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
457 * are we sure we do NOT LEAVE DOT FILES around? Especially on the
460 * smoker on k81 fetching from k75 to verify cascading works. See
461 2008-07-17 in upgradexxx and rsync-over-recentfile-3.pl.
463 * maybe the loop should wait for CHECKSUMS file after every upload. And
464 CPAN.pm needs to deal with timestamps in the future.
466 * do not forget the dirtymark!
468 Text: have a new flag on recentfiles with the meaning: if this
469 changes, you're required to run a full rsync over all the files. The
470 reason why we set it would probably be: some foul happened. we injected
471 files in arbitrary places or didn't inject them although they changed.
472 The content of the flag? Timestamp? The relation between the
473 recentfiles would have to be inheritance from the principal, because any
474 out of band changes would soon later propagate to the next recentfile.
476 By upping the flag often one can easily ruin the slaves.
478 last out of band change? dirtymark?
480 Anyway, this implies that we read a potentially existing recentfile
483 And it implies that we have an eventloop that keeps us busy in 2-3
484 cycles, one for current stuff (tight loop) and one for the recentfiles
485 (cascade when principal has changed), one for the old stuff after a
488 And it implies that the out-of-band change in any of the recentfiles
489 must have a lock on the principal file and there is the place to set the
492 * start a FAQ, especially quick start guide questions. Also to aid those
493 problematic areas where we have no good solution, like the "links"
496 * wish feedback when we are slow.
500 * Remove a few DEBUG statements.
502 * The multiple-rrr way of doing things needs a new option to rmirror,
503 like piecemeal or so. Not urgent because after the first pass through,
504 things run smoothely. It's only ugly during the first pass.
506 * I have the suspicion that the code is broken that decides if the
507 neighboring RF needs to be seeded. I fear when too much time has gone
508 between two calls (in our case more than one hour), it would not seed
509 the neighbor. Of course this will never be noticed, so we need a good
512 * local/localroot confusion: I currently pass both options but one must
515 * accounts for early birds on PAUSE rsync daemon.
517 * hardcoded 20 seconds
519 * who mirrors the index? DOING now.
521 * which CPAN mirrors offer rsync?
523 * visit all XXX, visit all _float places
525 * rename the pathdb stuff, it's too confusing. No idea how.
527 * rrr-inotify, backpan, rrr-register
529 2008-10-08 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
531 * current bugs: the pathdb seems to get no reset, the seeding of the
532 secondaryttl stuff seems not to have an effect. Have helped myself with
533 a rand(10), need to fix this back. So not checked in. Does the rand
536 The rand thing helps. The secondaryttl stuff was in the wrong line,
539 The pathdb stuff was because I called either _pathdb or __pathdb on the
540 wrong object. FIXED now.
542 * It's not so beautiful if we never fetch the recentfiles that are not
543 the principal, even if this is correct behaviour. We really do not need
544 them after we have fetched the whole content.
546 OK, we want a switch for that: secondaryttl DONE
548 2008-10-07 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
550 * bug: rrr-news --max does not count correctly. with "35" it shows me 35
551 lines but with 36 it shows 110. First it repeats 35, gives 70, and then
552 it lets 40 follow. FIXED
554 * See that the long running process really only updates the principal
555 file unless it has missed a timespan during which something happened. If
556 nothing happened, it must notice even when it misses the timespan. DONE
558 * we must throw away the pathdb when we have reached the end of Z. From
559 that moment we can have a very small pathdb because the only reason for
560 a pathdb is that we know to ignore old records in old files. We won't
561 need this pathdb again before the next full pass over the data is
562 necessary and then we will rebuild it as we go along. DONE
564 2008-10-06 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
566 * I think, Done::register_one is doing wrong in that it does not
567 conflate neighboring pieces. The covered() method cannot do this because
568 it has no recent_events array at hand. But register_one has it and could
569 do it and for some reason misses to do it (sometimes).
571 This means that the three tests I just wrote can probably not survive
572 because they test with an already broken Done structure.
574 The art now is to detect how it happens, then to reproduce, then write a
577 So from the logfile this is what happens: we have a good interval with
578 newest file being F1 at T1. Now remotely F1 gets a change and F2 goes on
579 top of it. Locally we now mirror F2 and open a new done interval for it.
580 Then we mirror F1 but this time with the timestamp T1b. And when we then
581 try to close the gap, we do not find T1 but instead something older. We
582 should gladly accept this older piece and this would fix this bug.
586 * bug to fix: when the 1h file changes while rmirror is running, we do
587 correctly sync the new files but never switch to the 6h file but rather
588 stay in a rather quick loop that fetches the 1h file again and again.
590 Is it possible that we initialize a new object? Or does
591 get_remote_recentfile_as_tempfile overwrite something in myself?
593 Want a new option: _runstatusfile => $file which frequently dumps the
594 state of all recentfiles to a file.
598 2008-10-04 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
600 * Todo: now teach update to verify the timestamp is about to write
601 against the previous and use _increase_a_bit if it doesn't comply with
602 strict monotony. DONE
604 * The problem of rounding. So far perl's default precision was
605 sufficient. One day it won't be. FakeFloat has an easy job when it is
606 only reading and other machines have written correctly. But when we want
607 to write a floating point number that is a bit larger than the other
608 one, then we need our own idea of precision.
610 Slaven said: just append a "1". This might be going towards the end of
611 usability too quickly. I'd like something that actually uses the decimal
612 system. Well, appending a 1 also does this but...
614 E.g. we have 1.0. nextup on this architecture is starting with
615 1.0000000000000004. So there is a gap to fill: 1,2,3. Now I have
616 taken the 1.0000000000000003 and the next user comes and the time tells
617 him 1.0 again. He has to beat my number without stepping over the
618 nextup. This is much less space than I had when I chose 1,2,3.
620 What is also irritating is that nextup is architecture dependent. The
621 128 bit guy must choose very long numbers to fit in between whereas the
622 other one with 16 bit uses larger steps. But then the algorithm is the
623 same for both, so that would be a nice thing.
625 I see two situation where we need this. One is when Time::HiRes returns
626 us a value that is <= the last entry in our recentfile. In this case
627 (let's call it the end-case) we must fill the region between that number
628 and the next higher native floating point number. The other is when we
629 inject an old file into an old recentfile (we would then also set a new
630 dirtymark). We find the integer value already taken and need a slightly
631 different one (let's call it the middle-case). The difference between
632 the two situations is that the next user will want to find something
633 higher than my number in the end-case and something lower than my number
636 So I suggest we give the function both a value and an upper bound and it
637 calculates us a primitive middle. The upper bound in the middle-case is
638 the next integer. The upper bound on the end-case is the nextup floating
639 point number. But the latter poses another problem: if we have occupied
640 the middle m between x and nextup(x), then the nextup(m) will probably
641 not be the same as nextup(x) because some rounding will take place
642 before the nextup is calculated and when the rounding reaches the
643 nextup(x), we will end up at nextup(nextup(x)).
645 So we really need to consider the nextup and the nextdown from there and
646 then the middle and that's the number we may approach asymptotically.
649 2008-10-03 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
651 * consider deprecating the use of RECENT.recent as a symlink. It turns
652 out to need extra hoops with the rsync options and just isn't worth it.
653 Or maybe these extra hoops are needed anyway for the rest of the tree?
654 Nope, can't be the case because not all filesystems support symlinks.
656 But before doing the large step, I'll deprecate the call of
657 get_remote_recentfile_as_tempfile with an argument. Rememberr this was
658 only introduced to resolve RECENT.recent and complicates the routine far
659 beyond what it deserves.
661 DONE. Won't deprecate RECENT.recent, just moved its handling to the
664 2008-10-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
666 * I think it's a bug that the rsync_option links must be set to true in
667 order to support RECENT.recent and that nobody cares to set it
668 automatically. Similar for ignore_link_stat_errors.
670 2008-09-27 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
672 * Todo: find all todos together and make a plan what is missing for a
675 - verifytree or something like that. fsck maybe.
677 - rersyncrecent, the script itself? What it do?
679 - a way to only mirror the recentfiles without mirroring the whole
680 remote system such that people can decide to mirror only partially see
681 also 2008-08-30. .shadow-xxx directory? this also needed for a
682 filesystem that is still incomplete and might need the mirrorfiles for
685 - long living objects that mirror again and again. Inject something
686 into ta, see how it goes over to tb.
688 - how do we continue filling up the DONE system when we use an object
689 for the second time? "fully covered" and "uptodate" or new terminology.
691 - overview called on the wrong file should be understandable
693 - the meta data field that must change when we fake something up so that
694 the downstream people know they have to re-fetch everything.
696 - how tolerant are we against missing files upstream? how do we keep
697 track? there are legitimate cases where we did read upstream index right
698 before a file got deleted there and then find that file as new and want
699 it. There are other cases that are not self healing and must be tracked
702 - how, exactly, do we have to deal with deletes? With rsync errors?
704 rsync: link_stat "/id/K/KA/KARMAN/Rose-HTMLx-Form-Related-0.07.meta" (in
705 authors) failed: No such file or directory (2)
707 The file above is a delete in 1h and a new in file 1M and the
708 delete in the locally running rmirror did not get propagated to the 1M
709 object. Bug. And the consequence is a standstill.
711 It seems that a slave that works with a file below the principal needs
712 to merge things all the way up to get rid of later deletes. Or keep
713 track of all deletes and skip them later. So we need a trackdeletes.pm
714 similar to the done.pm?
716 see also 2008-08-20 about spurious deletes that really have no add
717 counterpart and yet they are not wrong.
719 - consider the effect when resyncing the recentfile takes longer than
720 the time per loop. Then we never rsync any file. We need to diagnose
721 that and force an increase of that loop time. But when we later are fast
722 enough again because the net has recovered, then we need to switch back
723 to original parameters. ERm, no, it's enough to keep syncing at least
724 one file before refetching an index file.
726 - remember to verify that no temp files are left lying around and the
729 - status file for not long running jobs that want to track upstream with
732 - revisit all XXX _float areas and study Sub::Exporter DONE
734 - persistent DB even though we just said we do not need it. Just for
735 extended capabilities and time savings when, for example, upstream
736 announces a reset and we get new recentfiles and could then limit
737 ourselves to a subset of files (those that have a changed epoch) in a
738 first pass and would only then do the loop to verify the rest. Or
741 * Todo: aggregate files should know their feed and finding the principal
742 should be done stepwise. (?)
744 * Todo: DESTROY thing that unlocks. Today when I left the debuggerr I
745 left locks around. DONE
747 2008-09-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
749 * maybe extend the _overview so that it always says if and where the
750 last file is in the next file and where the next event in the next rf
751 would lie. No, don't like this anymore. REJECT
753 * take the two new redundant tests out again, only the third must
756 * Todo: add a sanity check if the merged structure is really pointing to
757 a different rf and that this different rf is larger. DONE
759 2008-09-25 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
761 * now test, if they are overlapping. And test if there is a file in the
762 next rf that would fit into this rf's interval.
764 1h 1222324012.8474 1222322541.7963 0.4086
765 6h 1222320411.2760 1222304207.6931 4.5010 missing overlap/gap!
766 1d 1222320411.2760 1222238750.5071 22.6835 large overlap
767 1W 1222313218.3626 1221708477.5829 167.9835
769 I suspect that somebody writes a merged timestamp without having merged
770 and then somebody else relies on it.
772 If aggregate is running, the intervals must not be extravagated, if it
773 is not running, there must not be bounds, the total number of events in
774 the system must be counted and must be controlled throughout the tests.
775 That the test required the additional update was probably nonsense,
776 because aggregate can cut pieces too. FIXED & DONE
778 2008-09-23 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
780 * rrr-aggregate seems to rewrite the RECENT file even if nothing has
783 2008-09-21 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
785 * Most apparent bug at the moment is that the recentfiles are fetched
786 too often. Only the principal should be fetched and if it has not
787 changed, the others should not be refetched. ATM I must admit that I'm
788 happy that we refetch more often than needed because I can more easily
789 fix bugs while the thing is running.
791 * Let's say, 1220474966.19501 is a timestamp of a file that is already
792 done but the done system does not know about it. The reason for the
793 failure is not known and we never reach the status uptodate because of
794 this. We must get over it.
796 Later it turns out that the origin server had a bug somewhere.
797 1220474966.19042 came after 1220474966.19501. Or better: it was in the
798 array of the recentfile one position above. The bug was my own.
800 2008-09-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
802 * There is the race condition where the server does a delete and the
803 slave does not yet know and then tries to download it because he sees
804 the new. So for this time window we must be more tolerant against
805 failure. If we cannot download a file, we should just skip it and should
806 not retry immediately. The whole system should discover the lost thing
807 later. Keeping track with the DONE system should really be a no brainer.
809 But there is something more: the whole filesystem is a database and the
810 recentfiles are one possible representation of it. It's a pretty useful
811 representation I think that's why I have implemented something around
812 it. But for strictly local operation it has little value. For local
813 operation we would much rather have a database. So we would enter every
814 recentfile reading and every rsync operation and for every file the last
815 state change and what it leads to. Then we would always ignore older
816 records without the efforts involved with recentfiles.
818 The database would have: path,recentepoch,rsyncedon,deletedon
820 Oh well, not yet clear where this leads to.
822 2008-09-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
824 * Bug: the bigloop ran into a funny endless loop after EWILHELM uploaded
825 Module-Build. It *only* rsynced the "1h" recentfile from that moment on.
827 * statusfile, maybe only on demand, alone to have a sharp debugging
828 tool. It is locked and all recentfiles dump themselves into it and we
829 can build a viewer that lets us know where we stand and what's inside.
831 * remember: only the principal recentfile needs expiration, all others
832 shall be expired by principal if it discovers that something has move
835 2008-09-18 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
837 * Always check if we stringify to a higher value than in the entry
840 * And in covered make an additional check if we would be able to see a
841 numerical difference between the two numbers and if we can't then switch
842 to a different, more expensive algorithm. Do not want to be caught by
843 floating surprises. DONE
845 2008-09-17 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
847 * caching has several aspects here: we can cache the interval of the
848 recentfile which only will change when the mtime of the file changes. We
849 must re-mirror the recentfile when its ttl has expired. Does have_read
850 tell you anything? It counts nothing at all. Only the mtime is
851 interesting. The ntuple mtime, low-epoch, high-epoch. And as a separate
852 thing the have_mirrored because it is unrelated to the mtime.
854 * Robustness of floating point calculations! I always thought that the
855 string calculated by the origin server for the floating representation
856 of the epoch time is just a string. When we convert it to a number and
857 later back to a string, the other computer might come to a different
858 conclusion. This must not happen, we want to preserve it under any
859 circumstances. I will have to write tests with overlong sequences that
860 get lost in arithmetic and must see if all still works well. DONE
862 But one fragile point remains: if one host considers a>b and the other
863 one considers them == but no eq. To prevent this, we must probably do
864 some extra homework. DONE
866 2008-09-16 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
868 * the concept of tracking DONE needs an object per recentfile that has
869 something like these methods:
871 do_we_have(xxx), we_have(xxx), do_we_have_all(xxx,yyy), reset()
873 covered() register() covered()
875 The unclear thing is how we translate points in time into intervals. We
876 could pass a reference to the current recent_events array when running
877 we_have(xxx) and let the DONE object iterate over it such that it only
878 has to store a list of intervals that can melt into each other. Ah, even
879 passing the list together with a list of indexes seems feasiable.
881 Or maybe ask for the inverted list?
883 Whenever the complete array is covered by the interval we say we are
884 fully covered and if the recentfile is not expired, we are uptodate.
886 2008-09-07 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
888 2008-09-05 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
890 * need a way to "return" the next entry after the end of a list. When
891 the caller says "before" or "after" we would like to know if he could
892 cover that interval/threshold or not because this influences the effect
893 of a newer timestamp of that recentfile. DONE with $opt{info}.
895 2008-09-04 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
897 * one of the next things to tackle: the equivalent of csync2 -TIXU.
899 loop implies tixu (?). Nope, something like --statefile decides. Per
902 T test, I init, X including removals, U nodirtymark
904 So we have no concept of dirtymarks, we only trust that since we are
905 running we have observed everything steadily. But people will not let
906 this program run forever so we must consider both startup penalty and
907 book keeping for later runs. We keep this for later. For now we write a
908 long running mirror that merges several intervals.
910 2008-09-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
912 * need to speed up the 02 test, it's not clever to sleep so much. Reduce
915 * rersyncrecent, the script: default to one week. The name of the switch
916 is --after. Other switches? --loop!
918 2008-08-30 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
920 * need a switch --skip-deletes (?)
922 * need a switch --enduser that tells us that the whole tempfile
923 discipline is not needed when there is no downstream user. (?)
925 Without this switch we cannot have a reasonable recent.pl that just
926 displays the recent additions. Either we accept to download everything.
927 Or we download temporary files without the typical rsync protocol
930 Or maybe the switch is --tmpdir? If --tmpdir would mean: do not use
931 File::Temp::tempdir, this might be a win.
933 2008-08-29 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
935 * apropos missing: we have no push, we never know the downstream
936 servers. People who know their downstream hosts and want to ascertain
937 something will want additional methods we have never thought about, like
938 update or delete a certain file.
940 2008-08-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
942 * tempted to refactor rmirror into resolve_symlink, localize, etc.
943 Curious if rsync_options=links equal 0 vs. 1 will make the expected
946 * rsync options: it's a bit of a pain that we usually need several rsync
947 options, like compress, links, times, checksum and that there is no
948 reasonable default except the original rsync default. I think wee can
949 safely assume that the rsync options are shared between all recentfile
950 instances within one recent tree.
952 2008-08-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
954 * deletes: if a delete follows an add quickly enough it may happen that
955 a downstream mirror did not see the add at all! It seems this needs to
956 be mentioned somewhere. The point here is that even if the downstream is
957 never missing the principal timeframe it may encounter a "delete" that
958 has no complimentary "add" anywhere.
960 2008-08-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
962 * I suspect the treat of metadata is incorrect during read or something.
963 The bug that I am watching is that between 06:08 and 06:09 the 6h file
964 contained more than 6 hours worth of data. At 06:08 we merged into the
965 1d file. We need to take snapshots of the 6h file over the course of an
966 hour or maybe only between XX:08 and XX:09? Nope, the latter is not
969 Much worse: watching the 1h file: right at the moment (at 06:35) it
970 covers 1218867584-1219120397 which is 70 hours.
972 Something terribly broken. BTW, 1218867584 corresponds to Sat Aug 16
973 08:19:44 2008, that is when I checked out last time, so it seems to be
974 aggregating and never truncating?
976 No, correct is: it is never truncating; but wrong is: it is aggregating.
977 It does receive a lot of events from time to time from a larger file.
978 Somehow a large file gets merged into the small one and because the
979 "meta/merged" attribute is missing, nobody is paying attention. I
980 believe that I can fix this by making sure that metadata are honoured
981 during read. DONE and test adjusted.
983 2008-08-17 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
985 * grand renaming plan
987 remotebase => remoteroot to fit well with localroot DONE
988 local_path() => localroot seems to me should already work DONE
989 recentfile_basename => rfilename no need to stress it has no slash DONE
991 filenameroot??? Doesn't seem too bad to me today. Maybe something like
992 kern? It would anyway need a deprecation cycle because it is an
993 important constructor.
995 * I like the portability that Data::Serializer brings us but the price
996 is that some day we might find out that it is slowing us a bit. We'll
999 2008-08-16 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1001 * should we not enter the interval of the principal (or the interval of
1002 the merging file?) in every aggregated/merged file?
1004 * we should aim at a first release and give up on thinking about
1005 sanitizing stuff and zloop. Let's just admit that a full traditional
1006 rsync is the only available sanitizer ATM. Otherwise it's complicated
1007 stuff: sanitizing on the origin server, sanitizing on the slaves,
1008 sanitizing forgotten files, broken timestamps, etc. Let's delay it and
1009 get the basics out before this becomes a major cause for mess.
1011 2008-08-13 Andreas Koenig <k@andreas-koenigs-computer.local>
1013 * On OSes not supporting symlinks we expect that RECENT.recent contains
1014 the contents of the principal recentfile. Actually this is identical on
1015 systems supporting symlinks. Simple, what follows from that is that we
1016 need to keep the serializer in the metadata because we cannot read it
1017 from the filename, doesn't it? Of course not. It's a chicken and egg
1018 problem. This leaves us with the problem to actually parse the
1019 serialized data to find out in which format it is. So who can do the 4
1020 or 5 magics we wanted to support? File::LibMagic?
1022 2008-08-09 Andreas Koenig <k@andreas-koenigs-computer.local>
1024 * remotebase and recentfile_basename are ugly names. Now that we need a
1025 word for the shortest/principal/driving recentfile too we should do
1028 localroot is good. rfile is good. local_path() is bad, local_path($path)
1029 is medium, filenameroot() is bad, remotebase is bad, recentfile is
1032 Up to now remotebase was the string that described the remote root
1033 directory in rsync notation, like pause.perl.org::authors. And
1034 recentfile_basename was "RECENT-1h.yaml".
1036 2008-08-08 Andreas Koenig <k@andreas-koenigs-computer.local>
1038 * The test that was added in today's checkin is a good start for a test
1039 of rmirror. We should have more methods in Recent.pm: verify,
1040 addmissingfiles. We should verify the current tree, then rmirror it and
1041 then verifytree the copy. We could then add some arbitrary file and let
1042 it be discovered by addmissingfiles, then rmirror again and then
1043 verifytree the copy again.
1045 Then we could start stealing from csync2 sqlite database [no port to
1046 OSX!] and fill a local DB. And methods to compare the database with the
1047 recentfiles. Our strength is that in principle we could maintain state
1048 with a single float. We have synced up to 1234567890.123456. If the Z
1049 file does not add new files all we have to do is mirror the new ones and
1052 This makes it clear that we should extend current protocol and declare
1053 that we cheat when we add files too late, just to help the other end
1054 keeping track. Ah yes, that's what was meant when zloop was mentioned
1057 Maybe need to revisit File::Mirror to help me with this task.
1059 2008-08-07 Andreas Koenig <k@andreas-koenigs-computer.local>
1061 * There must be an allow-me-to-truncate flag in every recentfile.
1062 Without it one could construct a sequence of updates winning the locking
1063 battle against the aggregator. Only if an aggregator has managed to
1064 merge data over to the next level, truncating can be allowed. DONE with
1067 2008-08-06 Andreas Koenig <k@andreas-koenigs-computer.local>
1069 * We should probably guarantee that no duplicates enter the aggregator
1072 2008-08-02 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1074 * To get merge operation faster would need a good benchmark test. What
1075 02 spits out isn't reliable enough and is dominated by many other
1078 commit 10176bf6b79865d4fe9f46e3857a3b8669fa7961
1079 Author: Andreas J. Koenig <k@k75.(none)>
1080 Date: Sat Aug 2 07:58:04 2008 +0200
1084 commit 3243120a0c120aaddcd9b1f4db6689ff12ed2523
1085 Author: Andreas J. Koenig <k@k75.(none)>
1086 Date: Sat Aug 2 11:40:29 2008 +0200
1088 there was a lot of trying but the effect is hardly measurable with
1091 * overhead of connecting seems high. When setting
1092 max_files_per_connection to 1 we see that.
1094 2008-08-01 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1096 * 1217622571.0889 - 1217597432.86734 = 25138.2215600014
1098 25138.2215600014/3600 = 6.98283932222261
1100 It jumps into the eye that this is ~ 7 hours, not ~6, so there seems to
1101 be a bug in the aggregator. FIXED
1103 2008-07-27 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1105 * e.g. id/Y/YE/YEWENBIN/Emacs-PDE-0.2.16.tar.gz: Do we have it, should
1106 we have it, can we mirror it, mirror it!
1108 I fear this needs a new class which might be called
1109 File::Rsync::Mirror::Recent. It would collect all recentfiles of a kind
1110 and treat them as an entity. I realize that a single recentfile may be
1111 sufficient for certain tasks and that it is handy for the low level
1112 programmer but it is not nice to use. If there is a delete in the 1h
1113 file then the 6h file still contains it. Seekers of the best information
1114 need to combine at least some of the recentfiles most of the time.
1116 There is the place for the Z loop!
1118 But the combination is something to collect in a database, isn't it. Did
1119 csync2 just harrumph?
1121 2008-07-26 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1123 * it just occurred to me that hosts in the same mirroring pool could
1124 help out each other even without rewriting the recentfile. Just fetch
1125 the stuff to mirror from several places, bingo. But that's something
1126 that should rather live in a separate package or in rsync directly.
1128 * cronjobs are unsuited because with ntp they would all come at the full
1129 minute and disturb each other. Besides that I'd hate to have a backbone
1130 with more than a few seconds latency.
1132 2008-07-25 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1134 * a second rsync server with access control for PAUSE. Port? 873 is the
1135 standard port, let's take 8873.
1137 * if there were a filesystem based on this, it would have a slow access
1138 to inexistent files. It would probably provide wrong readdir (only based
1139 on current content) or also a slow one (based on a recentfile written
1140 after the call). But it would provide fast access to existing files. Or
1141 one would deliberately allow slightly blurred answers based on some
1142 sqlite reflection of the recentfiles.
1144 * todo: write a variant of mirror() that combines two or more
1145 recentfiles and treats them like one
1147 * todo: signal handler to remove the tempfile
1149 2008-07-24 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1151 * now that we have the symlink I forgot how it should be used in
1154 * the z loop: add missing files to Z file. Just append them (instead of
1155 prepending). So one guy prepends something from the Y file from time to
1156 time and another guy appends something rather frequently. Collecting
1157 pond. When Y merges into Z, things get epoch and the collecting pond
1158 gets smaller. What exactly are "missing files"?
1160 take note of current epoch of the alpha file, let's call it the
1163 find all files on disk
1165 remove all files registered in the recentworld up to recent-ts
1167 remove all files that have been deleted after recent-ts according to
1170 2008-07-23 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1172 * rersyncrecent might be a cronjob with a (locked) state file which
1173 contains things like after and maybe last z sync or such?
1175 rrr-mirror might be an alternative name but how would we justify the
1176 three Rs when there is no Re-Rsync-Recent?
1178 With the --loop parameter it is an endless loop, without it is no loop.
1179 At least this is simple.
1181 * todo: new accssor z-interval specifies how often the Z file is updated
1182 against the filesystem. We probably want no epoch stamp on these
1183 entries. And we want to be able to filter the entries (e.g. no
1184 by-modules and by-category tree)
1186 2008-07-20 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1188 * Fill the Z file. gc or fsck or both. Somehow we must get the old files
1189 into Z. We do not need the other files filled up with filesystem
1192 * need interface to query for a file in order to NOT call update on
1193 PAUSE a second time within a short time.
1195 2008-07-19 Andreas J. Koenig <andreas.koenig.7os6VVqR@franz.ak.mind.de>
1197 * recommended update interval? Makes no sense, is different for
1204 change-log-default-name: "Todo"