Thread: Fwd: Re: ECPG error on inserting records to DB
>> I have problems with inserting records into DB by the C program that >> uses ECPG interface. I am using 7.1.2 in Mandrake 8.0. >> Sometimes it abnormally terminates with segmentation violation and >> sometimes it produces the following SQL error message: > >Could this be a memory problem? I am suspecting that ! I've noticed that behaviour when the system was uptime 41 days. After reboot (the whole system) all starts functioning properly. Question is who is responsible for that behaviour? A) my daemon process? B) kernel? C) postmaster? >The later means that ecpg gets an incorrect source, that is the text it >tries to scan is not correct. Does it print this during compile time (it >should)? No! These are runtime messages. >Could you strip down the example to a handy small program that you can send >me? No problem. -- Edward Pilipczuk
On Mon, Nov 26, 2001 at 05:03:30PM +0100, Edward Pilipczuk wrote: > >Could this be a memory problem? > > I am suspecting that ! > I've noticed that behaviour when the system was uptime 41 days. > After reboot (the whole system) all starts functioning properly. Sounds like memory too. :-) > Question is who is responsible for that behaviour? > A) my daemon process? Likely. > B) kernel? Very unlikely. > C) postmaster? Impossible. Not that the postmaster cannot have a memory leak but it shouldn't affect the client in such a way. Michael -- Michael Meskes Michael@Fam-Meskes.De Go SF 49ers! Go Rhein Fire! Use Debian GNU/Linux! Use PostgreSQL!
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 On 2001 November 27 06:53 am, Michael Meskes wrote: > > B) kernel? > > Very unlikely. Depends on which version of the kernel you're running and how much stress you're putting on the VM system. As you might recall, there was recently some changes and a bit of a mixup over the kernel's vm code. The old stuff in 2.2 is very stable and quite trustworthy. The newest stuff (2.4.14+) appears to be pretty stable too, but some of the stuff circa 2.4.9 has a less than perfect reputation, particularly under high loads. The ac series also has a stable vm system. As I don't use Mandrake, and you didn't explicitly mention your kernel version, I can't comment further, except to wish you luck. - -- Andrew G. Hammond mailto:drew@xyzzy.dhs.org http://xyzzy.dhs.org/~drew/ 56 2A 54 EF 19 C0 3B 43 72 69 5B E3 69 5B A1 1F 613-389-5481 5CD3 62B0 254B DEB1 86E0 8959 093E F70A B457 84B1 "To blow recursion you must first blow recur" -- me -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.0.6 (GNU/Linux) Comment: For info see http://www.gnupg.org iEYEARECAAYFAjwDjPgACgkQCT73CrRXhLG5KwCeIiaOQUppuI9DFoR+3/g1zQrY rR8An1mGfze3hn8+XdF6jFVsnpRSpBT2 =HURq -----END PGP SIGNATURE-----
On Tue, Nov 27, 2001 at 07:54:12AM -0500, Andrew G. Hammond wrote: > > > B) kernel? > > > > Very unlikely. > > Depends on which version of the kernel you're running and how much stress > you're putting on the VM system. As you might recall, there was recently > some changes and a bit of a mixup over the kernel's vm code. The old stuff Argh, I just forgot about that. Of course you're right. Thsi could be the reason. After all I wasn't able to tar 200MB with and earlier 2.4 release. > in 2.2 is very stable and quite trustworthy. The newest stuff (2.4.14+) > appears to be pretty stable too, but some of the stuff circa 2.4.9 has a less As far as VM is concerned yes, but I'd recommend not using .14 (several problems) and .15 (risk of data loss in umount). .16 seems to be okay. Michael -- Michael Meskes Michael@Fam-Meskes.De Go SF 49ers! Go Rhein Fire! Use Debian GNU/Linux! Use PostgreSQL!