Purging old Revisions
Is there an easy way to purge out the earliest 6,000 Revisions of the 9,600 that are in my repository? In a perfect world I would keep my revision numbers and timestamps, but that isn't 100% required.
Re: Purging old Revisions
On Wed, Nov 28, 2012 at 3:10 PM, Matthew Bluhm wrote: > > Is there an easy way to purge out the earliest 6,000 Revisions of the > 9,600 that are in my repository? > > In a perfect world I would keep my revision numbers and timestamps, but > that isn't 100% required. > > Short answer: No. Longer answer: It can be done, but everyone will have to check out new working copies and you might get weird results due to the loss of history. There is no "easy" way to do it, but if all you're doing is cutting off the first X revisions it's not the most complex task. You will keep the commit timestamps, but the revision numbers will be reset (unless you enter dummy revisions, which will then make your repository appear to own a modified DeLorean, unless you then further mangle those dummy revisions to change their timestamps). Version control systems are designed to keep your history. "Purging" goes against this core concept. Why are you embarking on this endeavor?
Re: Purging old Revisions
On Wed, Nov 28, 2012 at 2:36 PM, Andy Levy wrote: > >> Is there an easy way to purge out the earliest 6,000 Revisions of the >> 9,600 that are in my repository? >> >> In a perfect world I would keep my revision numbers and timestamps, but >> that isn't 100% required. >> > > Short answer: No. > > Longer answer: It can be done, but everyone will have to check out new > working copies and you might get weird results due to the loss of history. > There is no "easy" way to do it, but if all you're doing is cutting off the > first X revisions it's not the most complex task. You will keep the commit > timestamps, but the revision numbers will be reset (unless you enter dummy > revisions, which will then make your repository appear to own a modified > DeLorean, unless you then further mangle those dummy revisions to change > their timestamps). > > Version control systems are designed to keep your history. "Purging" goes > against this core concept. Why are you embarking on this endeavor? It's pretty much just a matter of time until someone does something that shouldn't have been done in any repository. Even if the answer is always going to be 'it can't be done', why question the motives of someone trying to fix things? -- Les Mikesell lesmikes...@gmail.com
Re: Purging old Revisions
Guten Tag Les Mikesell, am Mittwoch, 28. November 2012 um 21:42 schrieben Sie: > It's pretty much just a matter of time until someone does something > that shouldn't have been done in any repository. Even if the answer > is always going to be 'it can't be done', why question the motives of > someone trying to fix things? Because the reason for the question may be an accidently committed very important password or key to the world nobody should ever know of, but there already have been performance related question and/or some wanted to preserve disk space by only keeping most recent versions and in those cases it's more then questionable if the removable of versions will result in any benefit. In the scenarios I read in the past it wouldn't. Mit freundlichen Grüßen, Thorsten Schöning -- Thorsten Schöning E-Mail:thorsten.schoen...@am-soft.de AM-SoFT IT-Systeme http://www.AM-SoFT.de/ Telefon...05151- 9468- 55 Fax...05151- 9468- 88 Mobil..0178-8 9468- 04 AM-SoFT GmbH IT-Systeme, Brandenburger Str. 7c, 31789 Hameln AG Hannover HRB 207 694 - Geschäftsführer: Andreas Muchow
Re: Purging old Revisions
On Wed, Nov 28, 2012 at 2:50 PM, Thorsten Schöning wrote: > Guten Tag Les Mikesell, > am Mittwoch, 28. November 2012 um 21:42 schrieben Sie: > >> It's pretty much just a matter of time until someone does something >> that shouldn't have been done in any repository. Even if the answer >> is always going to be 'it can't be done', why question the motives of >> someone trying to fix things? > > Because the reason for the question may be an accidently committed > very important password or key to the world nobody should ever know > of, but there already have been performance related question and/or > some wanted to preserve disk space by only keeping most recent > versions and in those cases it's more then questionable if the > removable of versions will result in any benefit. In the scenarios I > read in the past it wouldn't. Does the answer change based on this? I'm pretty sure we have a mistakenly-committed DVD iso image or two AND some passwords that shouldn't be there. And some stuff imported from cvs in the wrong place, then moved, that I'd like to clean up so in the future those projects could be dumped and extracted separately. Starting over, I think I'd go one project per repository but at the time it seemed like access control would be cumbersome. -- Les Mikesell lesmikes...@gmail.com
Re: Purging old Revisions
On Wed, Nov 28, 2012 at 3:42 PM, Les Mikesell wrote: > On Wed, Nov 28, 2012 at 2:36 PM, Andy Levy wrote: > > > >> Is there an easy way to purge out the earliest 6,000 Revisions of the > >> 9,600 that are in my repository? > >> > >> In a perfect world I would keep my revision numbers and timestamps, but > >> that isn't 100% required. > >> > > > > Short answer: No. > > > > Longer answer: It can be done, but everyone will have to check out new > > working copies and you might get weird results due to the loss of > history. > > There is no "easy" way to do it, but if all you're doing is cutting off > the > > first X revisions it's not the most complex task. You will keep the > commit > > timestamps, but the revision numbers will be reset (unless you enter > dummy > > revisions, which will then make your repository appear to own a modified > > DeLorean, unless you then further mangle those dummy revisions to change > > their timestamps). > > > > Version control systems are designed to keep your history. "Purging" goes > > against this core concept. Why are you embarking on this endeavor? > > It's pretty much just a matter of time until someone does something > that shouldn't have been done in any repository. Even if the answer > is always going to be 'it can't be done', why question the motives of > someone trying to fix things? It's not so much about the motives as it is understanding why so that an appropriate answer can be provided. IOW, describe the *problem*, not the perceived solution. See also http://www.perlmonks.org/index.pl?node_id=542341
Re: Purging old Revisions
Guten Tag Les Mikesell, am Mittwoch, 28. November 2012 um 22:00 schrieben Sie: > Does the answer change based on this? One can maybe make more clear why removing versions may have not the desired effect at all and is therefore just a waste of time. Your accidently committed and never changed images are one example, but in the past there were questions were users tried to save disk space of files which get changed regularly but didn't knew of deltas, representation sharing etc. In both cases one would answer differently. Mit freundlichen Grüßen, Thorsten Schöning -- Thorsten Schöning E-Mail:thorsten.schoen...@am-soft.de AM-SoFT IT-Systeme http://www.AM-SoFT.de/ Telefon...05151- 9468- 55 Fax...05151- 9468- 88 Mobil..0178-8 9468- 04 AM-SoFT GmbH IT-Systeme, Brandenburger Str. 7c, 31789 Hameln AG Hannover HRB 207 694 - Geschäftsführer: Andreas Muchow
Re: Purging old Revisions
My repository is now over 10GB and it is rather cumbersome to take care of. There has been many mistakes such as committing extra files. The older revisions provide no value and only cause headaches. I initially thought that I could export out the repository at revision 6,000 then just apply a dump of transactions 6,001 to HEAD Thanks for your help On Wed, Nov 28, 2012 at 3:42 PM, Les Mikesell wrote: > On Wed, Nov 28, 2012 at 2:36 PM, Andy Levy wrote: > > > >> Is there an easy way to purge out the earliest 6,000 Revisions of the > >> 9,600 that are in my repository? > >> > >> In a perfect world I would keep my revision numbers and timestamps, but > >> that isn't 100% required. > >> > > > > Short answer: No. > > > > Longer answer: It can be done, but everyone will have to check out new > > working copies and you might get weird results due to the loss of > history. > > There is no "easy" way to do it, but if all you're doing is cutting off > the > > first X revisions it's not the most complex task. You will keep the > commit > > timestamps, but the revision numbers will be reset (unless you enter > dummy > > revisions, which will then make your repository appear to own a modified > > DeLorean, unless you then further mangle those dummy revisions to change > > their timestamps). > > > > Version control systems are designed to keep your history. "Purging" goes > > against this core concept. Why are you embarking on this endeavor? > > It's pretty much just a matter of time until someone does something > that shouldn't have been done in any repository. Even if the answer > is always going to be 'it can't be done', why question the motives of > someone trying to fix things? > > -- >Les Mikesell > lesmikes...@gmail.com >
Re: Purging old Revisions
On Wed, Nov 28, 2012 at 4:18 PM, Matthew Bluhm wrote: > My repository is now over 10GB and it is rather cumbersome to take care of. > > There has been many mistakes such as committing extra files. > > The older revisions provide no value and only cause headaches. > > I initially thought that I could export out the repository at revision > 6,000 then just apply a dump of transactions 6,001 to HEAD > > Do you truly have 6000 revisions - two thirds of your whole repository - which have NO value whatsoever? Probably not. Those revisions are the history of what you're producing. In the grand scheme of things, a 10GB repository isn't really that large. Rather than throw away *all* of your history, have you considered the possibility of removing just those revisions where large files (ISO images) which never should have been committed? This will preserve your history and reduce your repository size, and skip throwing the baby out with the bathwater. On Wed, Nov 28, 2012 at 3:42 PM, Les Mikesell wrote: > >> On Wed, Nov 28, 2012 at 2:36 PM, Andy Levy wrote: >> > >> >> Is there an easy way to purge out the earliest 6,000 Revisions of the >> >> 9,600 that are in my repository? >> >> >> >> In a perfect world I would keep my revision numbers and timestamps, but >> >> that isn't 100% required. >> >> >> > >> > Short answer: No. >> > >> > Longer answer: It can be done, but everyone will have to check out new >> > working copies and you might get weird results due to the loss of >> history. >> > There is no "easy" way to do it, but if all you're doing is cutting off >> the >> > first X revisions it's not the most complex task. You will keep the >> commit >> > timestamps, but the revision numbers will be reset (unless you enter >> dummy >> > revisions, which will then make your repository appear to own a modified >> > DeLorean, unless you then further mangle those dummy revisions to change >> > their timestamps). >> > >> > Version control systems are designed to keep your history. "Purging" >> goes >> > against this core concept. Why are you embarking on this endeavor? >> >> It's pretty much just a matter of time until someone does something >> that shouldn't have been done in any repository. Even if the answer >> is always going to be 'it can't be done', why question the motives of >> someone trying to fix things? >> >> -- >>Les Mikesell >> lesmikes...@gmail.com >> > >
Re: Purging old Revisions
On Wed, Nov 28, 2012 at 3:18 PM, Thorsten Schöning wrote: >> >> Does the answer change based on this? > > One can maybe make more clear why removing versions may have not the > desired effect at all and is therefore just a waste of time. Your > accidently committed and never changed images are one example, but in > the past there were questions were users tried to save disk space of > files which get changed regularly but didn't knew of deltas, > representation sharing etc. In both cases one would answer > differently. > OK, but I've probably got all of the above in one 60G+ repository. Is the best I can hope for that disk prices will keep going down (and speed up, in case I ever have to restore from a backup...)? -- Les Mikesell lesmikes...@gmail.com
Re: Purging old Revisions
There were some binary files that were included as part of the projects that weren't necessary, so its tons of seemingly small mistakes. The repository was started 5 1/2 years ago. Revision #6,000 was about 2 years ago. I have never used any history older than 24 months. Even though 10 GB doesn't seem big, about 2/3 of it is a waste, so its wasting time and money. I agree that the whole idea is too keep the history, but oldest transactions provide the least value for me. mb On Wed, Nov 28, 2012 at 4:28 PM, Andy Levy wrote: > > > On Wed, Nov 28, 2012 at 4:18 PM, Matthew Bluhm wrote: > >> My repository is now over 10GB and it is rather cumbersome to take care >> of. >> >> There has been many mistakes such as committing extra files. >> >> The older revisions provide no value and only cause headaches. >> >> I initially thought that I could export out the repository at revision >> 6,000 then just apply a dump of transactions 6,001 to HEAD >> >> > > > Do you truly have 6000 revisions - two thirds of your whole repository - > which have NO value whatsoever? Probably not. Those revisions are the > history of what you're producing. In the grand scheme of things, a 10GB > repository isn't really that large. > > Rather than throw away *all* of your history, have you considered the > possibility of removing just those revisions where large files (ISO images) > which never should have been committed? This will preserve your history and > reduce your repository size, and skip throwing the baby out with the > bathwater. > > > On Wed, Nov 28, 2012 at 3:42 PM, Les Mikesell wrote: >> >>> On Wed, Nov 28, 2012 at 2:36 PM, Andy Levy wrote: >>> > >>> >> Is there an easy way to purge out the earliest 6,000 Revisions of the >>> >> 9,600 that are in my repository? >>> >> >>> >> In a perfect world I would keep my revision numbers and timestamps, >>> but >>> >> that isn't 100% required. >>> >> >>> > >>> > Short answer: No. >>> > >>> > Longer answer: It can be done, but everyone will have to check out new >>> > working copies and you might get weird results due to the loss of >>> history. >>> > There is no "easy" way to do it, but if all you're doing is cutting >>> off the >>> > first X revisions it's not the most complex task. You will keep the >>> commit >>> > timestamps, but the revision numbers will be reset (unless you enter >>> dummy >>> > revisions, which will then make your repository appear to own a >>> modified >>> > DeLorean, unless you then further mangle those dummy revisions to >>> change >>> > their timestamps). >>> > >>> > Version control systems are designed to keep your history. "Purging" >>> goes >>> > against this core concept. Why are you embarking on this endeavor? >>> >>> It's pretty much just a matter of time until someone does something >>> that shouldn't have been done in any repository. Even if the answer >>> is always going to be 'it can't be done', why question the motives of >>> someone trying to fix things? >>> >>> -- >>>Les Mikesell >>> lesmikes...@gmail.com >>> >> >> >
Re: Purging old Revisions
Guten Tag Matthew Bluhm, am Mittwoch, 28. November 2012 um 23:57 schrieben Sie: > The repository was started 5 1/2 years ago. Which version does your repo has and which Subversion does your server use? Did you just upgrade your repo or made a full dump and load cycle with current versions of Subversion? New features like representation sharing could reduce the size of your repo a lot by preserving the history full history and working copies. > I agree that the whole idea is too keep the history, but oldest > transactions provide the least value for me. If you drop revisions, your working copies get useless and you have to checkout everything new. Just compare that time with whatever costs your currently maybe larger repo produces. You can easily compare the benefits of all approaches by copying your current repo and create a new one in the current format with dump/load and create a new with only the least revisions you want. Mit freundlichen Grüßen, Thorsten Schöning -- Thorsten Schöning E-Mail:thorsten.schoen...@am-soft.de AM-SoFT IT-Systeme http://www.AM-SoFT.de/ Telefon...05151- 9468- 55 Fax...05151- 9468- 88 Mobil..0178-8 9468- 04 AM-SoFT GmbH IT-Systeme, Brandenburger Str. 7c, 31789 Hameln AG Hannover HRB 207 694 - Geschäftsführer: Andreas Muchow
Re: Purging old Revisions
On Wed, Nov 28, 2012 at 5:57 PM, Matthew Bluhm wrote: > There were some binary files that were included as part of the projects that > weren't necessary, so its tons of seemingly small mistakes. > > The repository was started 5 1/2 years ago. Revision #6,000 was about 2 > years ago. I have never used any history older than 24 months. > > Even though 10 GB doesn't seem big, about 2/3 of it is a waste, so its > wasting time and money. > > I agree that the whole idea is too keep the history, but oldest transactions > provide the least value for me. This is an easier problem to solve than getting rid of old revisions. Just create a dump file of your repository and pipe it through svndumpfilter using the --exclude option to remove any paths you want to remove from your repository. Then load the end result into a new repository This will allow you to strip out the large binary files and recover disk space without needing to drop revisions. Simply removing the first 6000 revisions is not going to accomplish what you want. -- Thanks Mark Phippard http://markphip.blogspot.com/