[rrd-developers] problems with rrd_update() when updating rrd file
kevin brintnall
kbrint at rufus.net
Wed Jun 2 17:28:54 CEST 2010
On Wed, Jun 2, 2010 at 12:18 AM, Tobias Oetiker <tobi at oetiker.ch> wrote:
> Hi Kevin,
>
> Yesterday kevin brintnall wrote:
>
> > A couple questions for Tobi...
> >
> > (1) should rrd_update_r() call rrd_clear_error() at the top? Its return
> > call depends on rrd_test_error(). Would it make sense to
> save/clear/reset
> > the error? Not sure on this.
>
> well, the way the error flag is handled until now (theoretically)
> is, that whoever calls rrd_* functions has to rrd_clear_error prior to
> calling the function, and after calling the function the should
> check for errors ...
>
> this is modeled after errno(3)
> For some system calls and library functions (e.g.,
> getpriority(2)), -1 is a valid return on success. In such cases,
> a successful return can be distinguished from an error return by
> setting errno to zero before the call, and then, if the call
> returns a status that indicates that an error may have occurred,
> checking to see if errno has a non-zero value.
>
> I did a quick grep for this in program/src and find that it is
> mostly true except for a few cases. So the behaviour is quite
> consistant but only just ...
>
> as for rrd_update (and some others) doing stuff with rrd_test_error
> internally, I guess this is the real culprit as this is not in line
> with the behaviour you get from errno ...
>
> so I guess the parts of code that do use rrd_test_error internally
> should 'push' the prior state of rrd_error when they start working
> and at the very end 'pop' it if no error has been set.
>
> does this make sense ?
>
I understand what you're saying, but save/restore or push/pop would likely
complicate the code. As long as users of the API can think of it in terms
of errno, it's probably best to leave it as-is?
>
> > (2) Should we expose rrd_info_r ?
>
> aehm ... this is already exposed I think.
>
Quite right.. my mistake.
-kb
> cheers
> tobi
> >
> > -kb
> >
> > On Tue, Jun 1, 2010 at 7:42 AM, Tomáš Macek <maca02 at atlas.cz> wrote:
> >
> > > Hm, so rrd_clear_error() did the job. I inserted call of
> rrd_clear_error()
> > > before each rrd_*() calling and program works (see the line with
> > > "attention" comment). I've tried to simplify things as possible and
> here
> > > it is, maybe this will help to anyone else to solve his problems... Am
> I
> > > doing everything right now?
> > >
> > > Tomas
> > >
> > > -------------------------
> > >
> > > #include <stdio.h>
> > > #include <sys/types.h>
> > > #include <unistd.h>
> > > #include <stdlib.h>
> > > #include <stdarg.h>
> > > #include <rrd.h>
> > >
> > > void rrd_cmd(char *cmdstr) {
> > > char *strptr = NULL, *str = NULL;
> > > char *argv[64];
> > > int error = 0, i = 0;
> > > rrd_info_t *it = NULL;
> > > char rrdcmd[2048] = "";
> > >
> > > snprintf(rrdcmd, 2048, "%s", cmdstr);
> > >
> > > i = 0;
> > >
> > > /* memset update, create *argv[] */
> > > memset(argv, 0, 64 * sizeof(char *));
> > >
> > > for (strptr = rrdcmd; strptr != NULL;) {
> > > while ((str = strsep(&strptr, " ")) != NULL && str != '\0') {
> > > argv[i] = str;
> > > i++;
> > > }
> > > }
> > >
> > > rrd_clear_error(); /* attention!!! */
> > > it = rrd_info(i, argv);
> > > rrd_info_free(it);
> > > rrd_clear_error(); /* attention!!! */
> > > error = rrd_update(i, argv);
> > > if (error == -1) {
> > > printf("Error: RRD_UPDATE command: %s\n", rrd_get_error());
> > > rrd_clear_error();
> > > }
> > > }
> > >
> > > int update_rrd_cmd(char *rrd_name, char *ds_string, char
> *n_value_string,
> > > char *buff, int n) {
> > > memset(buff, 0, n);
> > > snprintf(buff, n, "update %s --template %s %s", rrd_name,
> ds_string,
> > > n_value_string);
> > > printf("%s: buff = %s\n", __FUNCTION__, buff);
> > > return 0;
> > > }
> > >
> > > int main(void) {
> > > char buff[2048] = "";
> > > update_rrd_cmd("./testfile.rrd", "ds0:ds1", "N:123:456", buff,
> 2048);
> > > rrd_cmd(buff);
> > > return 0;
> > > }
> > >
> > >
> > > On Tue, 01 Jun 2010 08:20:18 +0200, Tomáš Macek <maca02 at atlas.cz>
> wrote:
> > >
> > > > Hello again,
> > > > this problem is present since 1.4.0. I know it, because I've tried
> every
> > > > version since 1.3.8 (works), through 1.3.9 (works), 1.4.0 - 1.4.3
> does
> > > > not
> > > > work at all with syntax error as described below. Everytime I
> compiled
> > > > and
> > > > tried the same source code and linked with the proper library
> version.
> > > > What has changed in the API of rrd_update() in 1.4.0? I really don't
> > > > understand what I'm doing wrong. Any hints/help appreciated, looking
> at
> > > > the CHANGES file did not helped me.
> > > >
> > > > Tomas
> > > >
> > > >
> > > > On Thu, 20 May 2010 12:28:25 +0200, Tomáš Macek <maca02 at atlas.cz>
> wrote:
> > > >
> > > >> Hi, I have RHEL 5.5 with rrdtool 1.4.3 and I'm using it's API for
> > > >> updating
> > > >> rrd files. The problem is that when I pass arguments to the
> rrd_update()
> > > >> function, it always returns error saying "Usage: rrdtool update
> > > >> [--daemon
> > > >> <addr>] <file>". The same code worked really fine on 1.3.8. I looked
> at
> > > >> the CHANGELOG file and there was some line saying, that something
> has
> > > >> changed on checking optarg parameters.
> > > >>
> > > >> This is from gdb just before I pass the array into rrd_update()
> > > >> function:
> > > >> -----------
> > > >> Breakpoint 2, rrd_cmd (rrdcmd=0x634dc0 "update", cmd=1, di_cnt=2) at
> > > >> rrd_op.c:215
> > > >> 215 error = rrd_update(i, argv);
> > > >> (gdb) p argv
> > > >> $26 = {0x634dc0 "update", 0x634dc7 "cac_ap2aeth.rrd",
> > > >> 0x634e06 "--template", 0x634e11 "ds0:ds1", 0x634e19
> > > >> "N:2325779922:3582706806", 0x0 <repeats 59 times>}
> > > >> -----------
> > > >>
> > > >> and after program goes thourgh rrd_update(), the argv array looks
> like
> > > >> this:
> > > >> -----------
> > > >> 216 if (error == -1) {
> > > >> (gdb) p argv
> > > >> $28 = {0x634dc0 "update", 0x634e06 "--template", 0x634e11 "ds0:ds1",
> > > >> 0x634dc7 "cac_ap2aeth.rrd", 0x634e19 "N:2325779922:3582706806",
> > > >> 0x0 <repeats 59 times>}
> > > >> -----------
> > > >>
> > > >> Please, notice the replaced arguments after rrd_update()... I don't
> > > >> know,
> > > >> what I'm doing wrong, the code under 1.3.8 library worked just fine
> and
> > > >> according to the man rrdupdate the arguments are fine. Hope I'm not
> > > >> missing something stupid, but I was not able to find anything
> strange in
> > > >> rrd_update() function in the rrd_update.c.
> > > >> Let me know, if you need something more from me, thank you
> > > >>
> > > >> Regards, Tomas
> > > >>
> > > >> _______________________________________________
> > > >> rrd-developers mailing list
> > > >> rrd-developers at lists.oetiker.ch
> > > >> https://lists.oetiker.ch/cgi-bin/listinfo/rrd-developers
> > > >
> > > > _______________________________________________
> > > > rrd-developers mailing list
> > > > rrd-developers at lists.oetiker.ch
> > > > https://lists.oetiker.ch/cgi-bin/listinfo/rrd-developers
> > >
> > > _______________________________________________
> > > rrd-developers mailing list
> > > rrd-developers at lists.oetiker.ch
> > > https://lists.oetiker.ch/cgi-bin/listinfo/rrd-developers
> > >
> >
> >
> >
> >
>
> --
> Tobi Oetiker, OETIKER+PARTNER AG, Aarweg 15 CH-4600 Olten, Switzerland
> http://it.oetiker.ch tobi at oetiker.ch ++41 62 775 9902 / sb: -9900
--
kevin brintnall =~ /kbrint at rufus.net/
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.oetiker.ch/pipermail/rrd-developers/attachments/20100602/35b1bbe3/attachment.htm
More information about the rrd-developers
mailing list