I'm a little confused about what you are supposed to do when you discover a bug with CentOS (and by implication RHEL).
Is there some policy as to what to do and where to report the bug?
It doesn't seem appropriate to report it to http://bugs.centos.org because CentOS is kept in-sync with the upstream, so bugs will never get fixed in CentOS, only in the upstream distribution. Unless there is some feedback mechanism from this CentOS bugtracker to Redhat that I am unaware of.
So that makes the obvious place to report any bugs http://bugzilla.redhat.com, but do they (Redhat) have any opinion on people reporting bugs observed on a CentOS distribution there? Do they welcome it or do they discourage it? There is certainly no CentOS product classification in their bugzilla I can see.
So, in short, is there any CentOS bug reporting policy, and is it written down anywhere?
Hywel.
Hywel Richards wrote:
I'm a little confused about what you are supposed to do when you discover a bug with CentOS (and by implication RHEL).
Is there some policy as to what to do and where to report the bug?
It doesn't seem appropriate to report it to http://bugs.centos.org because CentOS is kept in-sync with the upstream, so bugs will never get fixed in CentOS, only in the upstream distribution. Unless there is some feedback mechanism from this CentOS bugtracker to Redhat that I am unaware of.
So that makes the obvious place to report any bugs http://bugzilla.redhat.com, but do they (Redhat) have any opinion on people reporting bugs observed on a CentOS distribution there? Do they welcome it or do they discourage it? There is certainly no CentOS product classification in their bugzilla I can see.
So, in short, is there any CentOS bug reporting policy, and is it written down anywhere?
Hywel.
I don't know if there is an official policy, but I report bugs at both and cross reference each to the other (unless they obviously only apply to one and not the other).
Red Hat does encourage the CentOS community to report bugs against their product, and in their "External Bugs" section there is even a selection for CentOS to enter the CentOS BugID number. Hence it makes sense to file against CentOS first, then file against RH and enter the External CentOS BigID number, and finally come back to CentOS and add the RH BugID for reference.
I agree it would be useful to have a written policy/guidance (if it doesn't already exist) - perhaps this is something we could get added to the Wiki.
Hywel Richards wrote:
Is there some policy as to what to do and where to report the bug?
report it on bugs.centos.org unless you have a RHEL subscription and are able to also reproduce the exact same issue under the exact same conditions, in which case you should report it at bugzilla.redhat.com
It doesn't seem appropriate to report it to http://bugs.centos.org because CentOS is kept in-sync with the upstream, so bugs will never get fixed in CentOS, only in the upstream distribution.
Thats not really 100% black/white situation - if there is a major issue that has implications, and we are able to fix it locally we will always do that and users are quite welcome to jump over and move into the contributors side of things for such issues.
However, the reason I said you should really report issues at bugs.centos.org is that someone needs to first make sure it is indeed not an issue introduced by the CentOS process ( We had had a few of those as well ), and the issue needs to go upstream.
- KB
It looks like maybe I was right to be confused as to where to report bugs as now there are two different answers.
Karanbir Singh wrote:
Hywel Richards wrote:
Is there some policy as to what to do and where to report the bug?
report it on bugs.centos.org unless you have a RHEL subscription and are able to also reproduce the exact same issue under the exact same conditions, in which case you should report it at bugzilla.redhat.com
Unfortunately I'm not in a position to do that - no RHEL subscription or installation. This is why previously I've been hesitant to report any problems there.
It doesn't seem appropriate to report it to http://bugs.centos.org because CentOS is kept in-sync with the upstream, so bugs will never get fixed in CentOS, only in the upstream distribution.
Thats not really 100% black/white situation - if there is a major issue that has implications, and we are able to fix it locally we will always do that and users are quite welcome to jump over and move into the contributors side of things for such issues.
The ones I've encountered so far are typically not major - not showshoppers for most people (otherwise they would get fixed a lot faster without any intervention from me anyway).
An example is this sprof problem that I reported in July: http://bugs.centos.org/view.php?id=3021 Another example is that the driver for my samsung printer prints garbage in high quality mode (the default), but works fine in standard mode (a useful thing to know).
However, the reason I said you should really report issues at bugs.centos.org is that someone needs to first make sure it is indeed not an issue introduced by the CentOS process ( We had had a few of those as well ), and the issue needs to go upstream.
This seems like a sensible policy to me. However, is there any process whereby bugs eventually get referred upstream, if not by the original bug reporter? It seems to me that the examples I gave above are unlikely to be introduced by CentOS, and should be appreciated by the upstream provider.
Whatever the system, I think it would be pretty useful to have something written down somewhere, so that CentOS users like me know what to do when they encounter bugs. E.g. for that sprof problem, I'm not sure if I should wait on CentOS developers to process it (in which case there isn't much to say), or whether I should be also reporting it to redhat, as suggested in the other response, in which case I think further instructions are necessary).
Hywel.
Hywel Richards wrote:
It looks like maybe I was right to be confused as to where to report bugs as now there are two different answers.
Karanbir Singh wrote:
Hywel Richards wrote:
Is there some policy as to what to do and where to report the bug?
report it on bugs.centos.org unless you have a RHEL subscription and are able to also reproduce the exact same issue under the exact same conditions, in which case you should report it at bugzilla.redhat.com
Unfortunately I'm not in a position to do that - no RHEL subscription or installation. This is why previously I've been hesitant to report any problems there.
I'm in the same boat, but have frequently jumped in on already-reported upstream bugs to confirm on CentOS, or asked other to see if the problem is in the upstream product for CentOS bugs.
It doesn't seem appropriate to report it to http://bugs.centos.org because CentOS is kept in-sync with the upstream, so bugs will never get fixed in CentOS, only in the upstream distribution.
Thats not really 100% black/white situation - if there is a major issue that has implications, and we are able to fix it locally we will always do that and users are quite welcome to jump over and move into the contributors side of things for such issues.
The ones I've encountered so far are typically not major - not showshoppers for most people (otherwise they would get fixed a lot faster without any intervention from me anyway).
An example is this sprof problem that I reported in July: http://bugs.centos.org/view.php?id=3021
Looks a lot like https://bugzilla.redhat.com/show_bug.cgi?id=458861 that is marked "Status: RELEASE_PENDING", so one might expect a fix in 5.3.
Added a note to the CentOS BZ and a link to CentOS to the upstream BZ.
Another example is that the driver for my samsung printer prints garbage in high quality mode (the default), but works fine in standard mode (a useful thing to know).
However, the reason I said you should really report issues at bugs.centos.org is that someone needs to first make sure it is indeed not an issue introduced by the CentOS process ( We had had a few of those as well ), and the issue needs to go upstream.
This seems like a sensible policy to me. However, is there any process whereby bugs eventually get referred upstream, if not by the original bug reporter? It seems to me that the examples I gave above are unlikely to be introduced by CentOS, and should be appreciated by the upstream provider.
Whatever the system, I think it would be pretty useful to have something written down somewhere, so that CentOS users like me know what to do when they encounter bugs. E.g. for that sprof problem, I'm not sure if I should wait on CentOS developers to process it (in which case there isn't much to say), or whether I should be also reporting it to redhat, as suggested in the other response, in which case I think further instructions are necessary).
Often CentOS developers will either report upstream or ask others to do so in the CentOS BZ notes. A clearly-stated stated policy would certainly help avoid confusion.
Phil
Phil Schaffner wrote:
Looks a lot like https://bugzilla.redhat.com/show_bug.cgi?id=458861 that is marked "Status: RELEASE_PENDING", so one might expect a fix in 5.3.
Added a note to the CentOS BZ and a link to CentOS to the upstream BZ.
Thanks Phil, that is good news indeed!
Hywel.
-
Hywel Richards -
Karanbir Singh -
Ned Slider -
Phil Schaffner