Re: rpm 5.x in Th
On Sep 23, 2012, at 6:45 PM, Łukasz Chrustek wrote: > >> I am trying to warn -- politely -- that you are in uncharted >> and unsupported waters if relying on --rollback as it used to >> exist in RPM. > > OK, understand. Now it isn't working anyway (and I think, that should > be removed - but it is Your project), so I need to be more carefull > (to not have the reason to use --rollbacke) and write some scripts now > to make my own --rollback. > Hint: if you asked -- nicely -- and gave me a reproducer I could likely repair whatever (likely modest) damage exists with --rollback. There's nothing in the rpm C implementation that isn't in the perl 1-liner, just untested. Or rip out --rollback in RPM if "false advertising" is the problem. Deleting code is utterly simple patching. > >> Hint: I release @rpm5.org (and run continuous integration >> in buildbots) with repackaging enabled. > >> Every distro I am aware of disables repackaging, and most >> user comments I have read suggest disabling to save disk space. > > I'm using repackege, and I think that in PLD there is more such > persons. Now using repackege (FOR ME) will change, but I will use it, > because I'm testing some new versions, and sometimes I don't have time > to finish tests. Then I ... yes - were using rollback to fast and > easly return to working version, next day/night I could return to > testing (this procedure were used by me _sometimes_ on production env > - after testing number of upgraded packeges by poldek -u -t package) > There are usage cases for repackaging no matter how/why --rollback is implemented. >> Sorry: I get gang-raped repeatedly by trolls. These >> days I have zero tolerance: >>You want a flamefest? Fine by me … > > I don't want/need flamefest. This 'piece' of You take to another > mailing list. > I don't want a flamefest either. Howvere, I am entirely in reactive mode: PLD chose when to upgrade to @rpm5.org, and bugs appear outside of my control. Hint: try launchpad.net/rpm bug reporting (and/or blueprints) if you wish to avoid flamefests. >> But yes this thread is a total waste of time trivially solved >> by some minor thought and scripting. > >> But --rollback in an RPM context is something other than a perl >> 1-liner. > > Yes. I can see. But, please do not involve some file system > transaction into it :) > The issues involved in logging 10-20 system calls are rather trivial compared to the issues of transactionally protecting (possibly buggy!) scriptlet operations. 73 de Jeff > -- > regards, > Łukasz Chrustek > > ___ > pld-devel-en mailing list > pld-devel-en@lists.pld-linux.org > http://lists.pld-linux.org/mailman/listinfo/pld-devel-en ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
Witam, > On Sep 23, 2012, at 6:13 PM, Łukasz Chrustek wrote: >> Hello, >> 3. "when the disk isn't/wasn't working properly, every solution is utterly useless" >> >>> You are clueless: saving state remotely permits an >>> entire machine to be recreated when hard drives fail. >> >>> Backups and off-site storage are well known remedies >>> for hard drive failures. >> >> You are writing about some strange (complicated and not easy to >> implement) solutions. I like the rollback behavior in old rpm, it was >> working in the way I expect. Now You, as author of rpm, are writing, >> that I'm only person in the world, which is using this option... If >> so, don't bother anymore, but leaving this option in --help is >> missleading for me, but as You stated earlier - only for me. >> > I am trying to warn -- politely -- that you are in uncharted > and unsupported waters if relying on --rollback as it used to > exist in RPM. OK, understand. Now it isn't working anyway (and I think, that should be removed - but it is Your project), so I need to be more carefull (to not have the reason to use --rollbacke) and write some scripts now to make my own --rollback. > Hint: I release @rpm5.org (and run continuous integration > in buildbots) with repackaging enabled. > Every distro I am aware of disables repackaging, and most > user comments I have read suggest disabling to save disk space. I'm using repackege, and I think that in PLD there is more such persons. Now using repackege (FOR ME) will change, but I will use it, because I'm testing some new versions, and sometimes I don't have time to finish tests. Then I ... yes - were using rollback to fast and easly return to working version, next day/night I could return to testing (this procedure were used by me _sometimes_ on production env - after testing number of upgraded packeges by poldek -u -t package) > Sorry: I get gang-raped repeatedly by trolls. These > days I have zero tolerance: > You want a flamefest? Fine by me … I don't want/need flamefest. This 'piece' of You take to another mailing list. > But yes this thread is a total waste of time trivially solved > by some minor thought and scripting. > But --rollback in an RPM context is something other than a perl > 1-liner. Yes. I can see. But, please do not involve some file system transaction into it :) -- regards, Łukasz Chrustek ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 6:32 PM, Tomasz Pala wrote: > On Sun, Sep 23, 2012 at 18:17:38 -0400, Jeffrey Johnson wrote: > >>> Try running replicated postgresql master node on failing drive and share >>> your results! >> >> Why? > > Just try it. > Pay me $100/hour and I will. >> The issue(s) involved with maintaining databases >> consistently with package manager upgrades are non-trivial >> to solve with a perl script like yours. > > My script does (well, should - it was 1-minute typing and might be > error-prone) exactly what shall be done. It doesn't alter any > database more than every other legitimate invocation, as it doesn't use > any hacks except reversing order of transactions performed. > In term of databases it's not rollback - but contrary to RDBMSes here this > doesn't matter; primary difference is that standalone database has > strictly specified and controlled I/O vectors, while rpm in properly > used system cannot track data, but metadata only; in *sql it's not normal > usage-scenario when someone replaces blob using filesystem tools, so the > assumption of data consistency is solid. > *sigh* You are so busy trolling me that you haven't bothered to look at what is implemented in RPM. For starters: 1) rpm-5.3.x+ embeds sqlite3 so that databases imports/exports are integrated with package management. There is no reason why postgresql cannot be implemented the same way, just that sqlite3 is a simpler API if/when attempting an embedding. 2) TPPM *uses* Berkeley DB ACID logs to extend to file system and scriptlet operations. Uts incoherent/inconsistent (and ignorant) to challeng me to upgrade a postgresql master server with packaging at the dame time you are claiming that a perl 1-liner is an adequate replacement for --rollback. >> So write a perl sc riot that permits upgrading postgresql >> masters now that you have solved --rollback with one line of perl. > > I ain't solved --rollback, just replaced it with something usable. > Good: You are still an ignorant troll. 73 de Jeff > -- > Tomasz Pala > ___ > pld-devel-en mailing list > pld-devel-en@lists.pld-linux.org > http://lists.pld-linux.org/mailman/listinfo/pld-devel-en ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sun, Sep 23, 2012 at 18:17:38 -0400, Jeffrey Johnson wrote: >> Try running replicated postgresql master node on failing drive and share >> your results! > > Why? Just try it. > The issue(s) involved with maintaining databases > consistently with package manager upgrades are non-trivial > to solve with a perl script like yours. My script does (well, should - it was 1-minute typing and might be error-prone) exactly what shall be done. It doesn't alter any database more than every other legitimate invocation, as it doesn't use any hacks except reversing order of transactions performed. In term of databases it's not rollback - but contrary to RDBMSes here this doesn't matter; primary difference is that standalone database has strictly specified and controlled I/O vectors, while rpm in properly used system cannot track data, but metadata only; in *sql it's not normal usage-scenario when someone replaces blob using filesystem tools, so the assumption of data consistency is solid. > So write a perl sc riot that permits upgrading postgresql > masters now that you have solved --rollback with one line of perl. I ain't solved --rollback, just replaced it with something usable. -- Tomasz Pala ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 6:18 PM, Tomasz Pala wrote: > In this episode... > > On Sun, Sep 23, 2012 at 18:03:06 -0400, Jeffrey Johnson wrote: > >>> he - wants - to - 'rollback' - using - his - repackages. ONLY >> >> troll++ > > ...Jeff-the-Omniscient knows better what user wanted to accomplish! (troll++)++ So don't use ACID-style rollback. Alternatively, >>> >>> Which one, this broken? >> >> troll++ > > ...Jeff-almighty can run broken code! > (troll++)++ You are a deluded idiot claiming that RPM --rollback is false advertising. >>> >>> Oh, so it was deliberately designed to break, right? >> >> troll++ >> >> Get a grip: PLD decided to upgrade to rpm-5.4.x. > > So is it broken or not? > Depends on whether you were born a bastard (or not). I do not know you genealogy sufficiently well to say whether it is broken or not ... >> Now you can all bugger off ? > > Last time I've checked we were on PLD maillist not rpm one, so if anyone > it's you who should go fuck himself. > troll++ 73 de Jeff > -- > Tomasz Pala > ___ > pld-devel-en mailing list > pld-devel-en@lists.pld-linux.org > http://lists.pld-linux.org/mailman/listinfo/pld-devel-en ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 6:13 PM, Łukasz Chrustek wrote: > Hello, > >>> 3. "when the disk isn't/wasn't working properly, every solution is utterly >>> useless" > >> You are clueless: saving state remotely permits an >> entire machine to be recreated when hard drives fail. > >> Backups and off-site storage are well known remedies >> for hard drive failures. > > You are writing about some strange (complicated and not easy to > implement) solutions. I like the rollback behavior in old rpm, it was > working in the way I expect. Now You, as author of rpm, are writing, > that I'm only person in the world, which is using this option... If > so, don't bother anymore, but leaving this option in --help is > missleading for me, but as You stated earlier - only for me. > I am trying to warn -- politely -- that you are in uncharted and unsupported waters if relying on --rollback as it used to exist in RPM. > Do You (other PLD users/devs) use repackege only with --oldpackage ? > Noone is/were using --rollback ? Or - You don't use repackage :P ? > You never regret installing new versions of some rpms :) ? > Hint: I release @rpm5.org (and run continuous integration in buildbots) with repackaging enabled. Every distro I am aware of disables repackaging, and most user comments I have read suggest disabling to save disk space. >>> 4. I've written this for this conversation as simplest demonstration of >>> what is missing. >>> > >> Yes: you are a simpleton. > > Jeff, I think You are going to far. Peace, men. > Sorry: I get gang-raped repeatedly by trolls. These days I have zero tolerance: You want a flamefest? Fine by me … But yes this thread is a total waste of time trivially solved by some minor thought and scripting. But --rollback in an RPM context is something other than a perl 1-liner. 73 de Jeff > -- > Pozdrawiam, > Łukasz Chrustek > > ___ > pld-devel-en mailing list > pld-devel-en@lists.pld-linux.org > http://lists.pld-linux.org/mailman/listinfo/pld-devel-en ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
In this episode... On Sun, Sep 23, 2012 at 18:03:06 -0400, Jeffrey Johnson wrote: >> he - wants - to - 'rollback' - using - his - repackages. ONLY > > troll++ ...Jeff-the-Omniscient knows better what user wanted to acomplish! >>> So don't use ACID-style rollback. Alternatively, >> >> Which one, this broken? > > troll++ ...Jeff-almighty can run broken code! >>> You are a deluded idiot claiming that RPM --rollback >>> is false advertising. >> >> Oh, so it was deliberately designed to break, right? > > troll++ > > Get a grip: PLD decided to upgrade to rpm-5.4.x. So is it broken or not? > Now you can all bugger off ? Last time I've checked we were on PLD maillist not rpm one, so if anyone it's you who should go fuck himself. -- Tomasz Pala ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 6:13 PM, Tomasz Pala wrote: > On Sun, Sep 23, 2012 at 17:56:55 -0400, Jeffrey Johnson wrote: > >> Which is what I said *repeatedly* in an attempt to adjust expectations of >> --rollback. Go read the Blablabla ... > > Whose expectations would you like to adjust? Seriously - you want _us_ to > expect rollback to work on global fs state? You must be an idiot then! The > question was: how to downgrade repackaged package set specifying the time > factor. > You insist on irrevelant shit. > troll++ >>> 3. "when the disk isn't/wasn't working properly, every solution is utterly >>> useless" >> >> You are clueless: saving state remotely permits an >> entire machine to be recreated when hard drives fail. > > Try running replicated postgresql master node on failing drive and share > your results! > Why? The issue(s) involved with maintaining databases consistently with package manager upgrades are non-trivial to solve with a perl script like yours. >> Backups and off-site storage are well known remedies >> for hard drive failures. > > Only when drive fails _after_ the backup was made. If my clock was set > properly during upgrades, my perl would select proper directories. > So write a perl sc riot that permits upgrading postgresql masters now that you have solved --rollback with one line of perl. Surely you can hack out a postgresql master software upgrade by next week instead of wasting time trolling me. 73 de Jeff > -- > Tomasz Pala > ___ > pld-devel-en mailing list > pld-devel-en@lists.pld-linux.org > http://lists.pld-linux.org/mailman/listinfo/pld-devel-en ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
Hello, >> 3. "when the disk isn't/wasn't working properly, every solution is utterly >> useless" > You are clueless: saving state remotely permits an > entire machine to be recreated when hard drives fail. > Backups and off-site storage are well known remedies > for hard drive failures. You are writing about some strange (complicated and not easy to implement) solutions. I like the rollback behavior in old rpm, it was working in the way I expect. Now You, as author of rpm, are writing, that I'm only person in the world, which is using this option... If so, don't bother anymore, but leaving this option in --help is missleading for me, but as You stated earlier - only for me. Do You (other PLD users/devs) use repackege only with --oldpackage ? Noone is/were using --rollback ? Or - You don't use repackage :P ? You never regret installing new versions of some rpms :) ? >> 4. I've written this for this conversation as simplest demonstration of what >> is missing. >> > Yes: you are a simpleton. Jeff, I think You are going to far. Peace, men. -- Pozdrawiam, Łukasz Chrustek ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sun, Sep 23, 2012 at 17:56:55 -0400, Jeffrey Johnson wrote: > Which is what I said *repeatedly* in an attempt to adjust expectations of > --rollback. Go read the Blablabla ... Whose expectations would you like to adjust? Seriously - you want _us_ to expect rollback to work on global fs state? You must be an idiot then! The question was: how to downgrade repackaged package set specifying the time factor. You insist on irrevelant shit. >> 3. "when the disk isn't/wasn't working properly, every solution is utterly >> useless" > > You are clueless: saving state remotely permits an > entire machine to be recreated when hard drives fail. Try running replicated postgresql master node on failing drive and share your results! > Backups and off-site storage are well known remedies > for hard drive failures. Only when drive fails _after_ the backup was made. If my clock was set properly during upgrades, my perl would select proper directories. -- Tomasz Pala ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 5:58 PM, Tomasz Pala wrote: > On Sun, Sep 23, 2012 at 17:45:08 -0400, Jeffrey Johnson wrote: > >>> The data is not gone - it exists in repackage. The _only_ thing that's >>> missing is some syntax sugar to ease downgrades. >> >> It doesn't exist in /var/spool/repackage if not enabled, >> or when bloat has been manually removed. > > Or when the earth is hit by a meteor. > That's not the case we're talking about. That's funny you still can't > understant, that: > > he - wants - to - 'rollback' - using - his - repackages. ONLY > troll++ >> So don't use ACID-style rollback. Alternatively, > > Which one, this broken? > troll++ >> You are a deluded idiot claiming that RPM --rollback >> is false advertising. > > Oh, so it was deliberately designed to break, right? > troll++ Get a grip: PLD decided to upgrade to rpm-5.4.x. Now you can all bugger off … 73 de Jeff > -- > Tomasz Pala > ___ > pld-devel-en mailing list > pld-devel-en@lists.pld-linux.org > http://lists.pld-linux.org/mailman/listinfo/pld-devel-en ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sun, Sep 23, 2012 at 23:59:48 +0200, Łukasz Chrustek wrote: >> cd /var/spool/repackage; ls | perl -ne 'use Time::Local; my >> ($mday,$mon,$year,$hour,$min,$sec) = split(/[\s.:]+/, "23.04.2012 >> 12:00:00"); my $time = >> timelocal($sec,$min,$hour,$mday,$mon-1,$year); print if $_>=$time' | >> sort -r | while read dir; do rpm -Uvh --oldpackage --force $dir/*; done > > > Thank You, Tomek, this 1-liner will be usefull with some more packages > :). Beware, as Jeff-the-Understandig said it will break if you have had your clock skewed! In contrary - his rpm will break always, so it's much more reliable! ;) -- Tomasz Pala ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
Hello, > cd /var/spool/repackage; ls | perl -ne 'use Time::Local; my > ($mday,$mon,$year,$hour,$min,$sec) = split(/[\s.:]+/, "23.04.2012 > 12:00:00"); my $time = > timelocal($sec,$min,$hour,$mday,$mon-1,$year); print if $_>=$time' | > sort -r | while read dir; do rpm -Uvh --oldpackage --force $dir/*; done Thank You, Tomek, this 1-liner will be usefull with some more packages :). -- Pozdrawiam, Łukasz Chrustek ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sun, Sep 23, 2012 at 17:45:08 -0400, Jeffrey Johnson wrote: >> The data is not gone - it exists in repackage. The _only_ thing that's >> missing is some syntax sugar to ease downgrades. > > It doesn't exist in /var/spool/repackage if not enabled, > or when bloat has been manually removed. Or when the earth is hit by a meteor. That's not the case we're talking about. That's funny you still can't understant, that: he - wants - to - 'rollback' - using - his - repackages. ONLY > So don't use ACID-style rollback. Alternatively, Which one, this broken? > You are a deluded idiot claiming that RPM --rollback > is false advertising. Oh, so it was deliberately designed to break, right? -- Tomasz Pala ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 5:50 PM, Tomasz Pala wrote: > On Sun, Sep 23, 2012 at 17:30:04 -0400, Jeffrey Johnson wrote: > >>> Thus noone expect this to work or even exists. >> >> What am I, chopped liver? >> >> I said exactly that I do not expect --rollback to Just Work, with additional >> history/context information regarding the history/rationale on a Mancoosi >> WP3 mailing list >> when the decision was made to _NOT_ continue with the previous >> implementation of --rollback. > > Blablabla - for the last time: I don't care. I don't know anyone who > cares. If you had no feedback for years, apparently no rpm user cares. > We are not talking about SUCH feature. If THIS doesn't work > /dev/null. > Which is what I said *repeatedly* in an attempt to adjust expectations of --rollback. Go read the Blablabla ... I do not understand your distinction. How is --rollback to be performed if operations are only partially reversed? >>> >>> cd /var/spool/repackage; ls | perl -ne 'use Time::Local; my >>> ($mday,$mon,$year,$hour,$min,$sec) = split(/[\s.:]+/, "23.04.2012 >>> 12:00:00"); my $time = timelocal($sec,$min,$hour,$mday,$mon-1,$year); print >>> if $_>=$time' | sort -r | while read dir; do rpm -Uvh --oldpackage --force >>> $dir/*; done >> >> Hint: Your script is utterly useless when the clock isn't/wasn't correct. > > 1. do you still not understand this distinction? troll++ > 2. do you know how package manager could restore packages now? troll++ > 3. "when the disk isn't/wasn't working properly, every solution is utterly > useless" You are clueless: saving state remotely permits an entire machine to be recreated when hard drives fail. Backups and off-site storage are well known remedies for hard drive failures. > 4. I've written this for this conversation as simplest demonstration of what > is missing. > Yes: you are a simpleton. 73 de Jeff > -- > Tomasz Pala > ___ > pld-devel-en mailing list > pld-devel-en@lists.pld-linux.org > http://lists.pld-linux.org/mailman/listinfo/pld-devel-en ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
Hello, >> I do not understand your distinction. >> >> How is --rollback to be performed if operations are >> only partially reversed? >> > i'm sure people want just to get old package back, to revert human > mistake of upgrading or some other reason for downgrade, because package > is misbehaving, not wanting perfect rollback like filesystem rollback. Yes, that was exacly the point. I simply forget to dump databases on testing env when upgrading postgres from 9.1 to 9.2. In this (rather simple) case I took the three repackaged rpms and did --oldpackage. With --rollback, I could earlier do this without searching rpms in /var/spool/repackage. > call it something else than "rollback", if it hurts your perfect world Or remove - if leaving this option would lead to making some magic with some filesystems :). -- Pozdrawiam, Łukasz Chrustek ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sun, Sep 23, 2012 at 17:30:04 -0400, Jeffrey Johnson wrote: >> Thus noone expect this to work or even exists. > > What am I, chopped liver? > > I said exactly that I do not expect --rollback to Just Work, with additional > history/context information regarding the history/rationale on a Mancoosi WP3 > mailing list > when the decision was made to _NOT_ continue with the previous > implementation of --rollback. Blablabla - for the last time: I don't care. I don't know anyone who cares. If you had no feedback for years, apparently no rpm user cares. We are not talking about SUCH feature. If THIS doesn't work > /dev/null. >>> I do not understand your distinction. >>> >>> How is --rollback to be performed if operations are >>> only partially reversed? >> >> cd /var/spool/repackage; ls | perl -ne 'use Time::Local; my >> ($mday,$mon,$year,$hour,$min,$sec) = split(/[\s.:]+/, "23.04.2012 >> 12:00:00"); my $time = timelocal($sec,$min,$hour,$mday,$mon-1,$year); print >> if $_>=$time' | sort -r | while read dir; do rpm -Uvh --oldpackage --force >> $dir/*; done > > Hint: Your script is utterly useless when the clock isn't/wasn't correct. 1. do you still not understand this distinction? 2. do you know how package manager could restore packages now? 3. "when the disk isn't/wasn't working properly, every solution is utterly useless" 4. I've written this for this conversation as simplest demonstration of what is missing. -- Tomasz Pala ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 5:40 PM, Tomasz Pala wrote: > On Sun, Sep 23, 2012 at 14:18:45 -0400, Jeffrey Johnson wrote: > >>> i'm sure people want just to get old package back, to revert human mistake >>> of upgrading or some other reason for downgrade, because package is >>> misbehaving, not wanting perfect rollback like filesystem rollback. >> >> RPM isn't responsible for human mistakes: no implementation can save users >> from mistakes: >> When the data is gone, you lose. >> This applies to erased files, removed packages, and dead disk drives. > > The data is not gone - it exists in repackage. The _only_ thing that's > missing is some syntax sugar to ease downgrades. > It doesn't exist in /var/spool/repackage if not enabled, or when bloat has been manually removed. >> Yes. You do realize I designed a >> "Transactionally Protected Package Management" >> to handle exactly and only package manager initiated operations? >> >> There is zero detectable interest several years later, measured by any of >> discussion or patch >> submission or attempts at using. > > What a surprise! - as I already told you, noone expects ACID-style > rollback from package manager, as this idea is broken by design. Rolling > back filesystem requires tool operating on filesystem (not application) > level. > So don't use ACID-style rollback. Alternatively, go honk your own Newer! Better! Bestest! implementation as you wish. >>> call it something else than "rollback", if it hurts your perfect world >> >> Call its whatever you want, rpm has been able to repackage existing >> content when erasing for most of this century. Users and distros are not >> enabling the functionality, and the RFE's for better continue incessantly. > > Apparently you completely don't understand this discussion. In short: > it is about using these repackages in comfortable way > Nothing more. No undoing triggers. No 'rollback'. Simple downgrade > _package set_ to the state at specified time. > This isn't a discussion, there is nothing to understand from ignorant trolls. >>> i my world, where i deploy software with rpm packages, i do poldek -u >>> package-old-version --downgrade as i do have old versions available in >>> package manager repository. but distro packages are not available that >>> easily, therefore people look into /var/spool/repackage dir >> >> So implement --rollback in poldek or yum or urpmi or apt or dpkg or smart or >> zypper >> or BTRFS or even the linux kernel if you wish. > > Indeed, this should be implemented in poldek. But rpm itself shouldn't > suggest having function, that doesn't and won't work - so remove this > 'rollback' > and don't confuse users. > You are a deluded idiot claiming that RPM --rollback is false advertising. 73 de Jeff ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sun, Sep 23, 2012 at 14:18:45 -0400, Jeffrey Johnson wrote: >> i'm sure people want just to get old package back, to revert human mistake >> of upgrading or some other reason for downgrade, because package is >> misbehaving, not wanting perfect rollback like filesystem rollback. > > RPM isn't responsible for human mistakes: no implementation can save users > from mistakes: > When the data is gone, you lose. > This applies to erased files, removed packages, and dead disk drives. The data is not gone - it exists in repackage. The _only_ thing that's missing is some syntax sugar to ease downgrades. > Yes. You do realize I designed a > "Transactionally Protected Package Management" > to handle exactly and only package manager initiated operations? > > There is zero detectable interest several years later, measured by any of > discussion or patch > submission or attempts at using. What a surprise! - as I already told you, noone expects ACID-style rollback from package manager, as this idea is broken by design. Rolling back filesystem requires tool operating on filesystem (not application) level. >> call it something else than "rollback", if it hurts your perfect world > > Call its whatever you want, rpm has been able to repackage existing > content when erasing for most of this century. Users and distros are not > enabling the functionality, and the RFE's for better continue incessantly. Apparently you completely don't understand this discussion. In short: it is about using these repackages in comfortable way Nothing more. No undoing triggers. No 'rollback'. Simple downgrade _package set_ to the state at specified time. >> i my world, where i deploy software with rpm packages, i do poldek -u >> package-old-version --downgrade as i do have old versions available in >> package manager repository. but distro packages are not available that >> easily, therefore people look into /var/spool/repackage dir > > So implement --rollback in poldek or yum or urpmi or apt or dpkg or smart or > zypper > or BTRFS or even the linux kernel if you wish. Indeed, this should be implemented in poldek. But rpm itself shouldn't suggest having function, that doesn't and won't work - so remove this 'rollback' and don't confuse users. -- Tomasz Pala ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 5:18 PM, Tomasz Pala wrote: > On Sun, Sep 23, 2012 at 11:48:53 -0400, Jeffrey Johnson wrote: > >> ? but triggers are executed as part of package management, >> changing file system state, and are not simply invertible. > > Thus noone expect this to work or even exists. > What am I, chopped liver? I said exactly that I do not expect --rollback to Just Work, with additional history/context information regarding the history/rationale on a Mancoosi WP3 mailing list when the decision was made to _NOT_ continue with the previous implementation of --rollback. >> I do not understand your distinction. >> >> How is --rollback to be performed if operations are >> only partially reversed? > > cd /var/spool/repackage; ls | perl -ne 'use Time::Local; my > ($mday,$mon,$year,$hour,$min,$sec) = split(/[\s.:]+/, "23.04.2012 12:00:00"); > my $time = timelocal($sec,$min,$hour,$mday,$mon-1,$year); print if $_>=$time' > | sort -r | while read dir; do rpm -Uvh --oldpackage --force $dir/*; done > Hint: Your script is utterly useless when the clock isn't/wasn't correct. 73 de Jeff ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sun, Sep 23, 2012 at 17:11:16 -0400, Jeffrey Johnson wrote: > I supplied accurate information regarding --rollback functionality > in this thread. Can't you understand that noone gives a shit about literal rollback? glen already told you - just rename this to --downgrade-with-available-rolled-back-packages-till-specified-timestamp and get rid of all the bloat. > Something that none of you understand at all. It's you who don't understand a _simple_ function that IS expected, instead trying to implement some complex feature doomed to be broken. Just like Hurd story. -- Tomasz Pala ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sun, Sep 23, 2012 at 11:07:51PM +0200, Tomasz Pala wrote: > >> > i'm sure people want just to get old package back > > rpm -Uvh --oldpackage, no? > Yeah, good luck after some heimdal/openldap-invoked update with > dozens of transactions and hundreds of rpm packages repackaged. Well I've actually got some good luck with apt-rpm pins but rather prefer a backup in the uncertain cases given that virtual machine/environment testing didn't veto it in the first place... If e.g. some databases get upgraded during the process then I don't get how anything but a snapshot/backup would help. Except for a stable magic wand but mine is just not there. :) -- WBR, Michael Shigorin -- Linux.Kiev http://www.linux.kiev.ua/ Sep 29, Kiev, Ukraine: -- http://conference.osdn.org.ua ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sun, Sep 23, 2012 at 11:48:53 -0400, Jeffrey Johnson wrote: > ? but triggers are executed as part of package management, > changing file system state, and are not simply invertible. Thus noone expect this to work or even exists. > I do not understand your distinction. > > How is --rollback to be performed if operations are > only partially reversed? cd /var/spool/repackage; ls | perl -ne 'use Time::Local; my ($mday,$mon,$year,$hour,$min,$sec) = split(/[\s.:]+/, "23.04.2012 12:00:00"); my $time = timelocal($sec,$min,$hour,$mday,$mon-1,$year); print if $_>=$time' | sort -r | while read dir; do rpm -Uvh --oldpackage --force $dir/*; done -- Tomasz Pala ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 4:49 PM, Michael Shigorin wrote: > > Guess it's rather a "choice of upstream" issue than a technical one. > So shop a a different upstream for --rollbackm … its _YOUR_ choice. I will leave the technical details researching how many "upstreams" have --rollback to your Google searching abilities. I predict that BTRFS! BTRFS! BTRFS! will be your answer. *shrug* 73 de Jeff ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 5:07 PM, Tomasz Pala wrote: > On Sun, Sep 23, 2012 at 23:49:52 +0300, Michael Shigorin wrote: > i'm sure people want just to get old package back >> >> rpm -Uvh --oldpackage, no? > > Yeah, good luck after some heimdal/openldap-invoked update with dozens > of transactions and hundreds of rpm packages repackaged. Of course one > can write a 1-liner sh to process this - just another thing Jeff doesn't > understand completely. > I supplied accurate information regarding --rollback functionality in this thread. Something that none of you understand at all. I will refrain from answering in the future; I'm not here to be trolled. 73 de Jeff ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sun, Sep 23, 2012 at 23:49:52 +0300, Michael Shigorin wrote: >> > i'm sure people want just to get old package back > > rpm -Uvh --oldpackage, no? Yeah, good luck after some heimdal/openldap-invoked update with dozens of transactions and hundreds of rpm packages repackaged. Of course one can write a 1-liner sh to process this - just another thing Jeff doesn't understand completely. -- Tomasz Pala ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sun, Sep 23, 2012 at 02:18:45PM -0400, Jeffrey Johnson wrote: > > i'm sure people want just to get old package back rpm -Uvh --oldpackage, no? > I'm not at all sure what people want, other than to complain. :] > Yes. You do realize I designed a > "Transactionally Protected Package Management" > to handle exactly and only package manager initiated operations? > > There is zero detectable interest several years later, measured > by any of discussion or patch submission or attempts at using. Guess it's rather a "choice of upstream" issue than a technical one. -- WBR, Michael Shigorin -- Linux.Kiev http://www.linux.kiev.ua/ Sep 29, Kiev, Ukraine: -- http://conference.osdn.org.ua ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 2:05 PM, Elan Ruusamäe wrote: > > i'm sure people want just to get old package back, to revert human mistake of > upgrading or some other reason for downgrade, because package is misbehaving, > not wanting perfect rollback like filesystem rollback. > RPM isn't responsible for human mistakes: no implementation can save users from mistakes: When the data is gone, you lose. This applies to erased files, removed packages, and dead disk drives. Packages don't misbehave, package monkeys make mistakes and distros don't do sufficient QA so users are affected by the errors. I'm not at all sure what people want, other than to complain. > rolling back filesystem state really assumes nothing else happens in your > filesystem than rpm packages. this is rarely true, there are logs, other > writable data that you expect not to be "rolled back" if you just downgrade > package. > Yes. You do realize I designed a "Transactionally Protected Package Management" to handle exactly and only package manager initiated operations? There is zero detectable interest several years later, measured by any of discussion or patch submission or attempts at using. > call it something else than "rollback", if it hurts your perfect world > Call its whatever you want, rpm has been able to repackage existing content when erasing for most of this century. Users and distros are not enabling the functionality, and the RFE's for better continue incessantly. > i my world, where i deploy software with rpm packages, i do poldek -u > package-old-version --downgrade as i do have old versions available in > package manager repository. but distro packages are not available that > easily, therefore people look into /var/spool/repackage dir > So implement --rollback in poldek or yum or urpmi or apt or dpkg or smart or zypper or BTRFS or even the linux kernel if you wish. 73 de Jeff ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On 23/09/12 18:48, Jeffrey Johnson wrote: On Sep 23, 2012, at 4:43 AM, Tomasz Pala wrote: > >Well, I might be wrong, but I think Lukasz expects something like >rpm -Uvh/var/spool/repackage/[date range reversed]*/* --force --nomd5 >only - i.e. not actual transactions rollback, but package set restore >(in proper order, thus preserving dependencies). > One might expect whatever outcome one wishes … \ >Restoring filesystem state (including things altered by triggers etc.) >is indeed dm/filesystem/backup software job and there's no point simulating >it on one more level. > … but triggers are executed as part of package management, changing file system state, and are not simply invertible. I do not understand your distinction. How is --rollback to be performed if operations are only partially reversed? i'm sure people want just to get old package back, to revert human mistake of upgrading or some other reason for downgrade, because package is misbehaving, not wanting perfect rollback like filesystem rollback. rolling back filesystem state really assumes nothing else happens in your filesystem than rpm packages. this is rarely true, there are logs, other writable data that you expect not to be "rolled back" if you just downgrade package. call it something else than "rollback", if it hurts your perfect world i my world, where i deploy software with rpm packages, i do poldek -u package-old-version --downgrade as i do have old versions available in package manager repository. but distro packages are not available that easily, therefore people look into /var/spool/repackage dir -- glen ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sep 23, 2012, at 4:43 AM, Tomasz Pala wrote: > > Well, I might be wrong, but I think Lukasz expects something like > rpm -Uvh /var/spool/repackage/[date range reversed]*/* --force --nomd5 > only - i.e. not actual transactions rollback, but package set restore > (in proper order, thus preserving dependencies). > One might expect whatever outcome one wishes … \ > Restoring filesystem state (including things altered by triggers etc.) > is indeed dm/filesystem/backup software job and there's no point simulating > it on one more level. > … but triggers are executed as part of package management, changing file system state, and are not simply invertible. I do not understand your distinction. How is --rollback to be performed if operations are only partially reversed? 73 de Jeff ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: [TH][patch] Network rc script pand friendly
In reply on message from Arkadiusz Miśkiewicz: DEFAULTHANDLING=yes That's wrong. It's per ifcfg config and not global. Ok, thanks for your advice. Anyone has some words to set NAP on system side ? blueman-projest is nice, but don't recorganise system-wide dns-masq config... -- Best Regards, Lord Blick ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: [TH][patch] Network rc script pand friendly
On Sunday 23 of September 2012, LordBlick wrote: > In reply on: > >> It allows starting empty pan0...panN interface for allowing any dhcpd > >> daemon to serve on it - bnepX appears as part of panX bridge. This patch > >> has no interference on other iterfaces, just other bridge name added. > >> Withoit it, at every update I need to patch it... > >> Please test and apply. Any comments also welcome. > > > > I'm sorry, previous file is incorect, reverse patch. This mail > > attachment is correct. > > Also its need to add some symlinks: > > # ln -s ifup-br /lib/rc-scripts/ifup-pan > > # ln -s ifdown-br /lib/rc-scripts/ifdown-pan > > Also its needed to set in /etc/sysconfig/network: > > DEFAULTHANDLING=yes That's wrong. It's per ifcfg config and not global. -- Arkadiusz Miśkiewicz, arekm / maven.pl ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: [TH][patch] Network rc script pand friendly
In reply on: It allows starting empty pan0...panN interface for allowing any dhcpd daemon to serve on it - bnepX appears as part of panX bridge. This patch has no interference on other iterfaces, just other bridge name added. Withoit it, at every update I need to patch it... Please test and apply. Any comments also welcome. I'm sorry, previous file is incorect, reverse patch. This mail attachment is correct. Also its need to add some symlinks: # ln -s ifup-br /lib/rc-scripts/ifup-pan # ln -s ifdown-br /lib/rc-scripts/ifdown-pan Also its needed to set in /etc/sysconfig/network: DEFAULTHANDLING=yes -- Best Regards, Lord Blick ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: [TH][patch] Network rc script pand friendly
W odpowiedzi na wiadomość z dnia 23.09.2012 11:21, od LordBlick: It allows starting empty pan0...panN interface for allowing any dhcpd daemon to serve on it - bnepX appears as part of panX bridge. This patch has no interference on other iterfaces, just other bridge name added. Withoit it, at every update I need to patch it... Please test and apply. Any comments also welcome. I'm sorry, previous file is incorect, reverse patch. This mail attachment is correct. Also its need to add some symlinks: # ln -s ifup-br /lib/rc-scripts/ifup-pan # ln -s ifdown-br /lib/rc-scripts/ifdown-pan -- Best Regards, Lord Blick --- /etc/rc.d/init.d/network2012-09-23 10:37:47.0 +0200 +++ /etc/rc.d/init.d/network.oryg 2012-04-04 08:25:55.0 +0200 @@ -144,7 +144,7 @@ interfaces_boot=` for i in $ifcfg_files; do case ${i##*/} in - ifcfg-lo|ifcfg-sit*|ifcfg-atm*|ifcfg-lec*|ifcfg-nas*|ifcfg-br*|ifcfg-macvlan*|ifcfg-macvtap*|ifcfg-*.*) continue ;; + ifcfg-lo|ifcfg-sit*|ifcfg-atm*|ifcfg-lec*|ifcfg-nas*|ifcfg-br*|ifcfg-pan*|ifcfg-macvlan*|ifcfg-macvtap*|ifcfg-*.*) continue ;; esac ONBOOT=""; USERS=""; . "$i" 2>/dev/null [ ${USERS:-no} != no ] && continue @@ -167,7 +167,7 @@ interfaces_br_boot=` for i in $ifcfg_files; do case ${i##*/} in - ifcfg-br*) ;; + ifcfg-br*|ifcfg-pan*) ;; *) continue ;; esac ONBOOT=""; USERS=""; . "$i" 2>/dev/null ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
[TH][patch] Network rc script pand friendly
It allows starting empty pan0...panN interface for allowing any dhcpd daemon to serve on it - bnepX appears as part of panX bridge. This patch has no interference on other iterfaces, just other bridge name added. Withoit it, at every update I need to patch it... Please test and apply. Any comments also welcome. -- Best Regards, Lord Blick --- /etc/rc.d/init.d/network2012-09-23 10:37:47.0 +0200 +++ /etc/rc.d/init.d/network.oryg 2012-04-04 08:25:55.0 +0200 @@ -144,7 +144,7 @@ interfaces_boot=` for i in $ifcfg_files; do case ${i##*/} in - ifcfg-lo|ifcfg-sit*|ifcfg-atm*|ifcfg-lec*|ifcfg-nas*|ifcfg-br*|ifcfg-pan*|ifcfg-macvlan*|ifcfg-macvtap*|ifcfg-*.*) continue ;; + ifcfg-lo|ifcfg-sit*|ifcfg-atm*|ifcfg-lec*|ifcfg-nas*|ifcfg-br*|ifcfg-macvlan*|ifcfg-macvtap*|ifcfg-*.*) continue ;; esac ONBOOT=""; USERS=""; . "$i" 2>/dev/null [ ${USERS:-no} != no ] && continue @@ -167,7 +167,7 @@ interfaces_br_boot=` for i in $ifcfg_files; do case ${i##*/} in - ifcfg-br*|ifcfg-pan*) ;; + ifcfg-br*) ;; *) continue ;; esac ONBOOT=""; USERS=""; . "$i" 2>/dev/null ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en
Re: rpm 5.x in Th
On Sat, Sep 22, 2012 at 16:21:22 -0400, Jeffrey Johnson wrote: > Noone is asking for --ropllback: why should I bother? I > certainly know what is/was intended and implemented. > > Adding logs for all file system content ends up > saving (at least) 3 copies of all content: Well, I might be wrong, but I think Lukasz expects something like rpm -Uvh /var/spool/repackage/[date range reversed]*/* --force --nomd5 only - i.e. not actual transactions rollback, but package set restore (in proper order, thus preserving dependencies). Restoring filesystem state (including things altered by triggers etc.) is indeed dm/filesystem/backup software job and there's no point simulating it on one more level. -- Tomasz Pala ___ pld-devel-en mailing list pld-devel-en@lists.pld-linux.org http://lists.pld-linux.org/mailman/listinfo/pld-devel-en