> First off - calm down. This is a voluntary list. Sorry, maybe you missunderstood me, I'm not upset in anyway and I'm very calm. I'm just trying to propose solutions that could help BackuPC beeing seen as a realy wonderfull solution. As for your comment, please discuss privately before shouting on a public list. sometimes, it can be a miss interpretation, and such comment will never help.
> Second, you didn't say that backing up the clients would annoy the users. > Now it's an issue? YES and this is a big YES: Why? because user never want backup (privately or at work). and there are many reasons for that. Privately, they never have time to do so, at work, backups means that they can be spyed. More over, when a backup occure, the disk activity is heavy and thus starting applications is significantly slowed. As a full backup can last upt to 1 hour, it's time consuming in their daily work. As it's their responsibility to accept being backed-up, they accept this drawback, but doubling backups will double that problem, and it's not acceptable. > Perhaps the users of those machines who have previously > lost data due to a poor (or no) backup regime would be more than happy to > be backed up to two servers. Thoses users had always a good backup regime (rsync to their home dirs on a file with snapshots and offsite storage). they never lost any files with this type of backup, and we are just seeking a new backup strategy to avoid consuming lots of tapes. From the user point of view, changing from one rsync to two rsync cannot be called progress. > Third, cool is irrelevant when you're talking about backups. Yes, but from the user point of view it is. Regarding 2 backup systems with equal safety, if one requires to backup twice the PC to be as safe as the other, then the user will label it as "not cool". > The word you're looking for is safe. Base your strategy on that. Of course > you're perfectly at liberty to call your own company's backup strategy > "stupid" but I know some of the full time Motorola backup admins and > "stupid" they aren't. No answer to that, please do it privatly. BTW, this is Not Motorola strategy, but our service provider company. As for the stragy, please let me explain as I think that you misunderstood me. Lest analyse the double strategy and then the single strategy with rsync to a remote site. With the double strategy you have the following constraints: - double backup (thus double time or more (see below)) - the 2 backups are not the same (client disk can change between both) - each files and it's copies from other PCs are transfered TWICE (not taking the benefit from single instance storage). - the second backup to the distant site may take ages and while altiris can continue an interrupted full backup, I've understood that BackupPC is not able to handle that and restarts from scrantch (correct me if I'm wrong). This may lead to the impossibility to do the second backup (to the distant site). (and during that time, the laptop must stay connected). - there is no way to verify that the 2 backups are ok. (impossible to compare sizes or checksums). - the 2 backups must occure during working hours. thus, you load your internet connection during time it is mainly needed by users. The only advantage is that it's easy to setup, safe in terms of algorithm and there are no technical showstoppers. With the single backup strategy and rsync to remote site: - only one local backup (no internet cnx load) - rsync to remote site can be done during night when users don't require internet bandwidth. - both backup are identical and can be compared in many ways (size, md5sum, transfer logs, ...) - only a compressed single instance of each file is tranfered reducing significantly internet transfer. - users PC hard disk activity (and connexion time requirement) is reduced to the minimum. The drawback is that: - it has to be tested. - rsync to the other site may be impossible to handle by the rsync system because of to much hard links to handle. (rsync, xfsdump, ...) - Technical issues explained by Les > Les is extremely experienced with backuppc and he's also suggested the > double backup scenario. There are no doubts on that, and I apreciated a lot the explanation. When I say that it is stubid to double backup, I'm not saying that Les or any sysadmin is stupid, I'm just saying that algorithmycaly speaking, it is avoidable to do 2 backups, unfortunately, it is not yet possible if I understand well Les. (More over, when I spoke to Les, I said that network is not a problem, but I should have said, network is not a problem outside working hours). Regards, Olivier. > > Damian > > Quoting [EMAIL PROTECTED]: > > Le Mardi 23 Août 2005 16:30, Damian O'Hara a écrit : > > > Hi Olivier, > > > > > > I know that another Motorola site in the UK has a backup server (not > > > backuppc :o) on each site that backs up both site's hosts. So all six > > > clients are: backed up by both backuppc servers. > > > > Yes I know those sites but I think this method is stupid as it has no > > benefit of the solution of rsync servers and it has the following > > disadvantages: - hosts are backed upo twice => twice more disk activity > > on the client and not > > cool for the user > > - if the file foo.doc is on 3 PCs in site A it'l be transferet 3 times to > > site > > B and only after that optimized by the pool which is stupid. > > > > > If your network is not a problem you could adopt that aproach. That way > > > you could recover any host data from any backup server. > > > Also, taking one backuppc server down won't affect the remaining > > > server. > > > > if site A crashes, then changing the config of archive hosts to normal > > host should be sufficient rieht? > > > > Olivier. > > > > > Damian > > > > > > Quoting [EMAIL PROTECTED]: > > > > Does this Disaster recovery safe cluster architecture possible with > > > > BackupPC? > > > > > > > > Let say we have 2 sites A and B > > > > Let says that we have 3 hosts per site to backup > > > > > > > > hosts 1A 2A 3A for site A and 1B 2B 3B for site B. > > > > > > > > If we declare hosts 1A 2A 3A plus hosts 1B 2B 3B as archive hosts on > > > > site A and at the same time > > > > we declare hosts 1B 2B 3B plus hosts 1A 2A 3A as archive hosts on > > > > site B > > > > > > > > then each night we to an "rsync-2.6.6 -H" (on /var/lib/backuppc/pc) > > > > between the 2 sites. > > > > > > > > Then we have nightly pool optimisation. > > > > > > > > Can this config work. If not why? > > > > If it can work, can we have the benefit to optimise redundant files > > > > between the 2 sites? > > > > > > > > If one server fail, is there a way to recovery host 1A on site A from > > > > B site? > > > > > > > > In normal operation, is it possible to download/recover files on host > > > > 1A from host 2B on site A? > > > > > > > > It such config could run (network is not a problem here), it would be > > > > realy cool as I wouldn't have to backup the backuppc server :-) > > > > > > > > -- > > > > Olivier LAHAYE > > > > Motorola Labs IT manager > > > > Saclay, FRANCE > > > > > > > > > > > > ------------------------------------------------------- > > > > SF.Net email is Sponsored by the Better Software Conference & EXPO > > > > September 19-22, 2005 * San Francisco, CA * Development Lifecycle > > > > Practices Agile & Plan-Driven Development * Managing Projects & Teams > > > > * Testing & QA Security * Process Improvement & Measurement * > > > > http://www.sqe.com/bsce5sf > > > > _______________________________________________ > > > > BackupPC-users mailing list > > > > [email protected] > > > > https://lists.sourceforge.net/lists/listinfo/backuppc-users > > > > http://backuppc.sourceforge.net/ > > > > > > ------------------------------------------------------- > > > SF.Net email is Sponsored by the Better Software Conference & EXPO > > > September 19-22, 2005 * San Francisco, CA * Development Lifecycle > > > Practices Agile & Plan-Driven Development * Managing Projects & Teams * > > > Testing & QA Security * Process Improvement & Measurement * > > > http://www.sqe.com/bsce5sf > > > _______________________________________________ > > > BackupPC-users mailing list > > > [email protected] > > > https://lists.sourceforge.net/lists/listinfo/backuppc-users > > > http://backuppc.sourceforge.net/ > > > > -- > > Olivier LAHAYE > > Motorola Labs IT manager > > Saclay, FRANCE > > > > > > ------------------------------------------------------- > > SF.Net email is Sponsored by the Better Software Conference & EXPO > > September 19-22, 2005 * San Francisco, CA * Development Lifecycle > > Practices Agile & Plan-Driven Development * Managing Projects & Teams * > > Testing & QA Security * Process Improvement & Measurement * > > http://www.sqe.com/bsce5sf > > _______________________________________________ > > BackupPC-users mailing list > > [email protected] > > https://lists.sourceforge.net/lists/listinfo/backuppc-users > > http://backuppc.sourceforge.net/ > > ------------------------------------------------------- > SF.Net email is Sponsored by the Better Software Conference & EXPO > September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices > Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA > Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf > _______________________________________________ > BackupPC-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/backuppc-users > http://backuppc.sourceforge.net/ -- POPI: MGBU -- Olivier LAHAYE CRM IT lab manager Saclay, FRANCE ------------------------------------------------------- SF.Net email is Sponsored by the Better Software Conference & EXPO September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf _______________________________________________ BackupPC-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/backuppc-users http://backuppc.sourceforge.net/
