Re: [sqlite] Strange Corruption
No, this is definitely not the reason in my case as I can reproduce this issue on every 3.7.2/3.7.3 machine I've tested after copying the database file (and only the database file) to these machines. Am 15.11.2010 15:41, schrieb Kirk Clemons: Not sure if it helps but I would see this quite frequently when an old journal file would be left behind in the same directory as the backup database. This could be why making a change to the database such as vacuum would prevent the corruption. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Strange Corruption
.dump (3.6.23.1) and afterwards sqlite3 /tmp/new.db /tmp/dump.sql with 3.7.2 worked and fixed the .backup problem (as expected). As I've already downgraded sqlite3 in our new firmware and patched the live-systems that were running with the new firmware I'll only have one machine to check whether the error comes back or not... So maybe I'll not be able to give feedback for a few weeks (as I can not enforce errors). However: if the error will not come back and could have to do with an already existing error in the database it would be quite interesting to know why integrity_check doesn't find the error before making a backup. (What means that a bug exists in any case: either in PRAGMA integrity_check or in the backup function). --- Pirmin Walthert Am 16.11.2010 14:22, schrieb Black, Michael (IS): Sorry, I meant .dump Given what you're describing I think it's worth finding out if you've found some bug in 3.7.2. The docs say 3.7.2 fixed a long-standing corruption bug. I don't know if that's related to this or not but sounds suspiciously close. So... #1 .dump the database #2 .import in into 3.7.2 #3 Run for a few days and see if you still get your backup problem. If still corrupt try 3.7.3 If it works then it sounds like the database was corrupt already and 3.7.2 just hits it. Michael D. Black Senior Scientist Advanced Analytics Directorate Northrop Grumman Information Systems From: sqlite-users-boun...@sqlite.org on behalf of Pirmin Walthert Sent: Tue 11/16/2010 7:09 AM To: sqlite-users@sqlite.org Subject: Re: [sqlite] EXTERNAL:Re: Strange Corruption Well indeed it wasn't 3.7.X that created the database originally. But it was always 3.7.2 that made the INSERTS/UPDATES that lead to the state in which the database couldn't be backed up anymore. So what do you mean in fact: 3.7.X maybe can't handle database structures created with older versions?! Even after doing a vacuum which fixed the bug I had the same errors again on the machines with 3.7.2 after a few days (after other INSERTS/UPDATES). About the thing I should test: There is no command called .export it seems?! But I think that I don't even have to test the thing you propose, as it will work almost for sure = like already stated several times one little tiny tiny tiny change already fixes the error. As an .export/.import will change some bits for sure this will already change the situation! Am 16.11.2010 13:53, schrieb Black, Michael (IS): I thought of another test you should try. Do an .export of your original database using 3.6.23.1 and .import it (constructing a new database). Then try your backup. If that works then you're just seeing corruption in the original database that 3.6.23.1 handles (since it created it). If it doesn't work import into 3.7.3 and test backup again. If it doesn't work then try cutting the SQL in half until it does work. Maybe you'll finally get a small enough size you can post. Michael D. Black Senior Scientist Advanced Analytics Directorate Northrop Grumman Information Systems From: sqlite-users-boun...@sqlite.org on behalf of Pirmin Walthert Sent: Tue 11/16/2010 6:27 AM To: sqlite-users@sqlite.org Subject: EXTERNAL:Re: [sqlite] Strange Corruption No, this is definitely not the reason in my case as I can reproduce this issue on every 3.7.2/3.7.3 machine I've tested after copying the database file (and only the database file) to these machines. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Strange Corruption
Not sure if it helps but I would see this quite frequently when an old journal file would be left behind in the same directory as the backup database. This could be why making a change to the database such as vacuum would prevent the corruption. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Strange Corruption
No, this is definitely not the reason in my case as I can reproduce this issue on every 3.7.2/3.7.3 machine I've tested after copying the database file (and only the database file) to these machines. Am 15.11.2010 15:41, schrieb Kirk Clemons: Not sure if it helps but I would see this quite frequently when an old journal file would be left behind in the same directory as the backup database. This could be why making a change to the database such as vacuum would prevent the corruption. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Strange Corruption
Sorry, I meant .dump Given what you're describing I think it's worth finding out if you've found some bug in 3.7.2. The docs say 3.7.2 fixed a long-standing corruption bug. I don't know if that's related to this or not but sounds suspiciously close. So... #1 .dump the database #2 .import in into 3.7.2 #3 Run for a few days and see if you still get your backup problem. If still corrupt try 3.7.3 If it works then it sounds like the database was corrupt already and 3.7.2 just hits it. Michael D. Black Senior Scientist Advanced Analytics Directorate Northrop Grumman Information Systems From: sqlite-users-boun...@sqlite.org on behalf of Pirmin Walthert Sent: Tue 11/16/2010 7:09 AM To: sqlite-users@sqlite.org Subject: Re: [sqlite] EXTERNAL:Re: Strange Corruption Well indeed it wasn't 3.7.X that created the database originally. But it was always 3.7.2 that made the INSERTS/UPDATES that lead to the state in which the database couldn't be backed up anymore. So what do you mean in fact: 3.7.X maybe can't handle database structures created with older versions?! Even after doing a vacuum which fixed the bug I had the same errors again on the machines with 3.7.2 after a few days (after other INSERTS/UPDATES). About the thing I should test: There is no command called .export it seems?! But I think that I don't even have to test the thing you propose, as it will work almost for sure = like already stated several times one little tiny tiny tiny change already fixes the error. As an .export/.import will change some bits for sure this will already change the situation! Am 16.11.2010 13:53, schrieb Black, Michael (IS): I thought of another test you should try. Do an .export of your original database using 3.6.23.1 and .import it (constructing a new database). Then try your backup. If that works then you're just seeing corruption in the original database that 3.6.23.1 handles (since it created it). If it doesn't work import into 3.7.3 and test backup again. If it doesn't work then try cutting the SQL in half until it does work. Maybe you'll finally get a small enough size you can post. Michael D. Black Senior Scientist Advanced Analytics Directorate Northrop Grumman Information Systems From: sqlite-users-boun...@sqlite.org on behalf of Pirmin Walthert Sent: Tue 11/16/2010 6:27 AM To: sqlite-users@sqlite.org Subject: EXTERNAL:Re: [sqlite] Strange Corruption No, this is definitely not the reason in my case as I can reproduce this issue on every 3.7.2/3.7.3 machine I've tested after copying the database file (and only the database file) to these machines. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
Re: [sqlite] Strange Corruption
.dump (3.6.23.1) and afterwards sqlite3 /tmp/new.db /tmp/dump.sql with 3.7.2 worked and fixed the .backup problem (as expected). As I've already downgraded sqlite3 in our new firmware and patched the live-systems that were running with the new firmware I'll only have one machine to check whether the error comes back or not... So maybe I'll not be able to give feedback for a few weeks (as I can not enforce errors). However: if the error will not come back and could have to do with an already existing error in the database it would be quite interesting to know why integrity_check doesn't find the error before making a backup. (What means that a bug exists in any case: either in PRAGMA integrity_check or in the backup function). --- Pirmin Walthert Am 16.11.2010 14:22, schrieb Black, Michael (IS): Sorry, I meant .dump Given what you're describing I think it's worth finding out if you've found some bug in 3.7.2. The docs say 3.7.2 fixed a long-standing corruption bug. I don't know if that's related to this or not but sounds suspiciously close. So... #1 .dump the database #2 .import in into 3.7.2 #3 Run for a few days and see if you still get your backup problem. If still corrupt try 3.7.3 If it works then it sounds like the database was corrupt already and 3.7.2 just hits it. Michael D. Black Senior Scientist Advanced Analytics Directorate Northrop Grumman Information Systems From:sqlite-users-boun...@sqlite.org on behalf of Pirmin Walthert Sent: Tue 11/16/2010 7:09 AM To:sqlite-users@sqlite.org Subject: Re: [sqlite] EXTERNAL:Re: Strange Corruption Well indeed it wasn't 3.7.X that created the database originally. But it was always 3.7.2 that made the INSERTS/UPDATES that lead to the state in which the database couldn't be backed up anymore. So what do you mean in fact: 3.7.X maybe can't handle database structures created with older versions?! Even after doing a vacuum which fixed the bug I had the same errors again on the machines with 3.7.2 after a few days (after other INSERTS/UPDATES). About the thing I should test: There is no command called .export it seems?! But I think that I don't even have to test the thing you propose, as it will work almost for sure = like already stated several times one little tiny tiny tiny change already fixes the error. As an .export/.import will change some bits for sure this will already change the situation! Am 16.11.2010 13:53, schrieb Black, Michael (IS): I thought of another test you should try. Do an .export of your original database using 3.6.23.1 and .import it (constructing a new database). Then try your backup. If that works then you're just seeing corruption in the original database that 3.6.23.1 handles (since it created it). If it doesn't work import into 3.7.3 and test backup again. If it doesn't work then try cutting the SQL in half until it does work. Maybe you'll finally get a small enough size you can post. Michael D. Black Senior Scientist Advanced Analytics Directorate Northrop Grumman Information Systems From:sqlite-users-boun...@sqlite.org on behalf of Pirmin Walthert Sent: Tue 11/16/2010 6:27 AM To:sqlite-users@sqlite.org Subject: EXTERNAL:Re: [sqlite] Strange Corruption No, this is definitely not the reason in my case as I can reproduce this issue on every 3.7.2/3.7.3 machine I've tested after copying the database file (and only the database file) to these machines. ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users ___ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
