O > > Always remember that a psotgres data base on the harddisk can be > > manipulated accidentally / maliciously without postgres even running. > > These are the cases where you need row level CRCs. > > "There is no security without physical security." If somebody can > change the row contents, they can also change the row and/or block > checksum to match. They may, but in a proper setup they won't be able to access the CRC log files. That way, you can still detect alterations. I presume anyway that most alterations would be rather accidental than malicious, and in that case the CRC is extremely helpful Horst
- Re: CRC was: Re: [HACKERS] beta testing version Nathan Myers
- Re: CRC was: Re: [HACKERS] beta testing versi... Tom Lane
- Re: CRC was: Re: [HACKERS] beta testing v... Nathan Myers
- [HACKERS] RFC: CRC datatype Horst Herb
- RE: [HACKERS] RFC: CRC datatype Christopher Kings-Lynne
- Re: [HACKERS] RFC: CRC datatype Horst Herb
- Re: [HACKERS] RFC: CRC datatype Tom Lane
- Re: [HACKERS] RFC: CRC datat... Horst Herb
- Re: [HACKERS] RFC: CRC datat... Tom Lane
- Re: [HACKERS] RFC: CRC datatype Nathan Myers
- Re: CRC was: Re: [HACKERS] b... Horst Herb
- Re: CRC was: Re: [HACKERS] beta testing versi... Bruce Guenter
- Re: CRC was: Re: [HACKERS] beta testing v... Ian Lance Taylor
- Re: CRC was: Re: [HACKERS] beta testi... Bruce Guenter
- Re: CRC was: Re: [HACKERS] beta testing v... Tom Lane
- Re: CRC was: Re: [HACKERS] beta testi... Bruce Guenter
- Re: CRC was: Re: [HACKERS] beta t... Tom Lane
- Re: CRC was: Re: [HACKERS] b... Bruce Guenter
- [HACKERS] Re: CRC Nathan Myers
- Re: [HACKERS] Re: CRC Bruce Guenter
- Re: [HACKERS] Re: CRC Tom Lane