Hi Maik, thanks for your fast response :) I have read your Ideas For Projects and it all looks very good, also very ambitious... and I cannot say that I would have enough time to do the whole job at this point, but would be willing to help out with some coding.
I did, in my one day working with it, already find a coding error in pgsql-archlog which resulted in the program exiting due to freeing the same memory twice. I have fixed that function and will send a pull request if you are interested in fixes to this code. Keeping in mind that I haven't gotten it to successfully back up yet.. still getting zero size backups.. I know both C and python so have no major religious problem with choosing to do it over in python, and it is certainly a good choice for portability but I must admit I like to have fewer dependencies and C has less in this situation. Also being in the position here where we really need to get the backups working because we have insufficient backups here, so the idea of improving the existing code.. incremental improvement.. and getting it working for us asap is probably more realistic right now so far as what my workplace will spare of my time on this project. But to take 100% responsibility for it is probably not going to be possible and not even wise given that I am still learning both bareos and the postgres backup API capabilities. Perhaps if I were to figure out the issue I am having, get the module working for me, and make some improvements it would be sufficient for you to take it into your regular package process. Right now I am still trying understand the difference between what is proposed vs existing pgsql-plugin. It seems that the existing pgsql-plugin in :db mode does a backup of the postgres data directory (I think that /etc/postgresql should be included with this as well), and in :wal mode it backs up the wal files which have been staged by pgsql-archlog. So that's the incremental backup. So, if I understand the intentions (I can only guess a little bit, because I have not gotten it totally working yet), it is already similar to at least the first part of what is suggested in the ideas for projects?? But I would take your word for it if you tell me that this plugin is actually not very good and it's better to start over. One thing I wondered, both working with the existing plugin and reading your ideas, was the idea of having a continuous backup where the wal files are backed up immediately, as soon as they are completed, without needing to trigger timed backup jobs. Since postgres executes the archiver when needed. Also it seems to me that there is no need to make any copy of the wal file since it is ok to tell postgres (via the archiver exit code) that the file was not archived yet.. so when postgres tells the archiver to archive a particular file, it triggers the backup process on that file, and then returns a non-zero status. This will cause postgres to retry after an interval, and if our backup has completed by that time, the archiver can return 0 and postgres will delete the wal file at that time. I see you are talking about Barman integration in your ideas and I just discovered that system and started looking at it. We'd like everything to be backed up by the same system and that means Bareos but the possibility exists to use Barman just for the databases. Ah, in a final note.. I see that the code I compiled.. which I am using on the client, is version 16, whereas the director and sd are from version 15 packages.. and maybe that's why the backups are coming out as 0 size? Anyway thanks again for your thoughts and your efforts, they are much appreciated here. Ben On Tuesday, February 23, 2016 at 5:04:27 AM UTC-5, maik.aussendorf wrote: > Hello Ben, > > we've included and adapted the source code of the pgsql plugin into > our repository but we do not support it. That is one of the reasons, > why there are no packages. > > We have a PGSQL support on our "long-term" list and we also have > ideas to replace the existing code by a Python plugin, which makes > use of the Postgres backup interface, an outline is here: > > https://www.bareos.org/en/ideas-for-projects.html > > However, don't expect that there will be much progress in the next > time regarding PGSQL, unless > > - someone from the community picks up that subject > > or > > - someone funds further development > > > > Another option is to talk about consultancy for Postgres backups > with our partners or Bareos. > > > > Just let me know, if you are open for one of those options. > > > > Regards > > Maik > > > > > > > Am 23.02.2016 um 05:52 schrieb Light > O'Matic: > > > > > > > > > > > > > > > > Ok.. so I figured out how to compile it... > > > And it seems to be working.. the archiver is coping the WAL > files into /tmp, I see entries in pgsql_archivelogs with > status = 3, and I am able to status the client from > bconsole. I have set up the two jobs (:wal and :db). > > > > > What is happening is that the jobs show "Success" status, but > everything is 0 bytes (no data being saved). > > > > > > > > > > > > > > On Mon, Feb 22, 2016 at 5:05 PM, Ben <[email protected]> > wrote: > > Hi. > > > > First time Bareos user here... > > I have got Bareos 15.2.2 installed on Debian jessie using > binary packages from bareos.org, > and it seems to be working well. > > > > Now I want to set up backup jobs for my postgres servers > (not the Bareos Catalog but my production data), and I see > contrib-psql-plugin, which does exactly what I want.. so I > want to use it :) > > > > I cannot find a binary package for this.. but searching > around, I see that some people in the last year have > compiled it from source with success, so I get the source: > > > > git clone https://github.com/bareos/bareos.git > > git clone https://github.com/bareos/contrib-pgsql-plugin.git > > > > First, I built bareos, with success. > > > > Then, following instructions found at > https://github.com/inteos/pgsql-plugin/wiki/Install > , I went to compile the psql plugin. > > > > I copied Makefile.Linux to Makefile > > Then went to edit Makefile, and noticed that everything in > there still references Bacula and Bacula directories. > > > > I understand the history and that Bareos is the open fork of > Bacula.. so it's not the references to Bacula that concern > me, but the number of lines of the Makefile I would have to > change to get it to build. > > > > Am I using the wrong source for the psql plugin? > > Am I going about this the wrong way? > > Are there binary packages I am not aware of? > > Has this plugin simply fallen behind and is not useable at > the moment without more work? > > > > In short, please set me straight as to the status of the > psql-plugin with Bareos.. where should I be starting? Any > hints appreciated! > > > > Thanks very much > > Ben > > > > -- > > You received this message because you are subscribed to > a topic in the Google Groups "bareos-users" group. > > To unsubscribe from this topic, visit > https://groups.google.com/d/topic/bareos-users/VcA0FMa4wS4/unsubscribe. > > To unsubscribe from this group and all its topics, send > an email to [email protected]. > > To post to this group, send email to > [email protected]. > > For more options, visit https://groups.google.com/d/optout. > > > > > > > -- > > You received this message because you are subscribed to the Google > Groups "bareos-users" group. > > To unsubscribe from this group and stop receiving emails from it, > send an email to [email protected]. > > To post to this group, send email to [email protected]. > > For more options, visit https://groups.google.com/d/optout. > > > > > -- > With kind regards // Mit freundlichen Grüßen > -- > Maik Außendorf [email protected] > Bareos GmbH & Co. KG Phone: +49221630693-93 > http://www.bareos.com Fax: +49221630693-10 > > Sitz der Gesellschaft: Köln | Amtsgericht Köln: HRA 29646 > Komplementär: Bareos Verwaltungs-GmbH > Geschäftsführer: Stephan Dühr, M. Außendorf, > J. Steffens, P. Storz, M. v. Wieringen -- You received this message because you are subscribed to the Google Groups "bareos-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. To post to this group, send email to [email protected]. For more options, visit https://groups.google.com/d/optout.
