renayama19661014 at ybb
Nov 27, 2011, 4:17 PM
Post #8 of 8
Re: [Patch]Remove unnecessary loop handling of data_directory for postfix.
[In reply to]
About the second patch which I contributed, how do you think?
--- On Mon, 2011/11/21, renayama19661014 [at] ybb <renayama19661014 [at] ybb> wrote:
> Hi Raoul,
> Thank you for comment.
> Because postfix check did not give back the details to a result as for RA, I recognized that the details of the log were necessary.
> I changed a check of data_directory.
> And I abolish a suggestion street in front, the loop.
> This is because the plural setting is not admitted because it added a check.
> Please please confirm my correction.
> And please commit a correction.
> Best Regards,
> Hideo Yamauchi.
> --- On Sat, 2011/11/19, Raoul Bhatia [IPAX] <r.bhatia [at] ipax> wrote:
> > Hi Hideo-san!
> > On 2011-11-16 11:36, renayama19661014 [at] ybb wrote:
> > > I think that the same check has been already carried out in a resource agent.
> > >
> > > (snip)
> > > # run Postfix internal check, if not probing
> > > if ! ocf_is_probe; then
> > > $binary $OPTIONS check>/dev/null 2>&1
> > > ret=$?
> > > if [ $ret -ne 0 ]; then
> > > ocf_log err "Postfix 'check' failed." $ret
> > > return $OCF_ERR_GENERIC
> > > fi
> > > fi
> > > (snip)
> > >
> > >
> > > That means, after all is not the loop check of data_directory unnecessary?
> > postfix check is called after all other checks have passed and, you're
> > right, it also checks the required directories.
> > i think i had some issues though:
> > > # check spool/queue and data directories (if applicable)
> > > # this is required because "postfix check" does not catch all errors
> > but i cannot remember the exact problems anymore.
> > anyways, postfix check will return a "OCF_ERR_GENERIC"
> > which is regarded as a soft error (!)  and will
> > a. not hint the user or a gui application to the exact problem and
> > b. will lead to a restart of the failed resource on the same node
> > the more in-depth check will fail with OCF_ERR_INSTALLED  or
> > OCF_ERR_PERM  and will
> > c. give more information in this regard and
> > d. migrates the resource to a different node,
> > which makes sense if i.e. the "shared" queue directory (nfs, etc.)
> > isn't available.
> > i think that this behavior is good and checking the most commonly
> > modified directories separately has been very helpful in my setups.
> > but of course, i'm open for comments.
> > > #Sorry...Because English is weak, I may understand your opinion by mistake.
> > no worries. english isn't my first language either and until now we
> > managed to work things out, right? :)
> > cheers,
> > raoul
> >  http://www.linux-ha.org/doc/dev-guides/_literal_ocf_err_generic_literal_1.html
> >  http://www.linux-ha.org/doc/dev-guides/_literal_ocf_err_installed_literal_5.html
> >  http://www.linux-ha.org/doc/dev-guides/_literal_ocf_err_perm_literal_4.html
> > -- ____________________________________________________________________
> > DI (FH) Raoul Bhatia M.Sc. email. r.bhatia [at] ipax
> > Technischer Leiter
> > IPAX - Aloy Bhatia Hava OG web. http://www.ipax.at
> > Barawitzkagasse 10/2/2/11 email. office [at] ipax
> > 1190 Wien tel. +43 1 3670030
> > FN 277995t HG Wien fax. +43 1 3670030 15
> > ____________________________________________________________________
Linux-HA-Dev: Linux-HA-Dev [at] lists
Home Page: http://linux-ha.org/