From atumball at redhat.com Wed Nov 7 05:39:35 2012 From: atumball at redhat.com (Amar Tumballi) Date: Wed, 07 Nov 2012 11:09:35 +0530 Subject: [sos-devel] Gluster component update for sos In-Reply-To: <5086D9EC.8070102@redhat.com> References: <5086D9EC.8070102@redhat.com> Message-ID: <5099F417.80004@redhat.com> On 10/23/2012 11:24 PM, Amar Tumballi wrote: > Hi all, > > I am attaching two patches which we are planing to have as part of sos > for RHS2.0 (based on sos-2.2_17.el6_2). > > 0001-* was the first patch, done to perform 'statedump' only on one > machine and to avoid 'gluster volume statedump'/'gluster volume status > *' commands from the gluster component, with which we saw couple of issues. > > We had couple of comments from Bryn on this patch, and the result is > 0002-* patch which tried to remove the usage of /tmp/* directories. Also > fixed an issue which Bryn saw on RHEL machines (without glusterd binary). > Guys, any review comments on the patches? would like to get it in. Regards, Amar From adam.stokes at canonical.com Wed Nov 7 20:42:34 2012 From: adam.stokes at canonical.com (Adam Stokes) Date: Wed, 07 Nov 2012 15:42:34 -0500 Subject: [sos-devel] Gluster component update for sos In-Reply-To: <5099F417.80004@redhat.com> References: <5086D9EC.8070102@redhat.com> <5099F417.80004@redhat.com> Message-ID: <509AC7BA.90704@canonical.com> On 11/7/2012 12:39 AM, Amar Tumballi wrote: > On 10/23/2012 11:24 PM, Amar Tumballi wrote: >> Hi all, >> >> I am attaching two patches which we are planing to have as part of sos >> for RHS2.0 (based on sos-2.2_17.el6_2). >> >> 0001-* was the first patch, done to perform 'statedump' only on one >> machine and to avoid 'gluster volume statedump'/'gluster volume status >> *' commands from the gluster component, with which we saw couple of >> issues. >> >> We had couple of comments from Bryn on this patch, and the result is >> 0002-* patch which tried to remove the usage of /tmp/* directories. Also >> fixed an issue which Bryn saw on RHEL machines (without glusterd >> binary). >> > > Guys, any review comments on the patches? would like to get it in. > > Regards, > Amar > > _______________________________________________ > Sos-devel mailing list > Sos-devel at redhat.com > https://www.redhat.com/mailman/listinfo/sos-devel > The easiest way to get these reviewed and pulled in would be to fork the project on github and create your pull requests that way. Unfortunately I do not have admin priv's to the master repo or I'd take a look at it Thanks Adam From bmr at redhat.com Thu Nov 8 09:38:29 2012 From: bmr at redhat.com (Bryn M. Reeves) Date: Thu, 08 Nov 2012 09:38:29 +0000 Subject: [sos-devel] Gluster component update for sos In-Reply-To: <509AC7BA.90704@canonical.com> References: <5086D9EC.8070102@redhat.com> <5099F417.80004@redhat.com> <509AC7BA.90704@canonical.com> Message-ID: <509B7D95.3020707@redhat.com> On 07/11/12 20:42, Adam Stokes wrote: > On 11/7/2012 12:39 AM, Amar Tumballi wrote: >> On 10/23/2012 11:24 PM, Amar Tumballi wrote: >>> Hi all, >>> >>> I am attaching two patches which we are planing to have as part of sos >>> for RHS2.0 (based on sos-2.2_17.el6_2). >>> >>> 0001-* was the first patch, done to perform 'statedump' only on one >>> machine and to avoid 'gluster volume statedump'/'gluster volume status >>> *' commands from the gluster component, with which we saw couple of >>> issues. >>> >>> We had couple of comments from Bryn on this patch, and the result is >>> 0002-* patch which tried to remove the usage of /tmp/* directories. Also >>> fixed an issue which Bryn saw on RHEL machines (without glusterd >>> binary). >>> >> >> Guys, any review comments on the patches? would like to get it in. >> >> Regards, >> Amar >> >> _______________________________________________ >> Sos-devel mailing list >> Sos-devel at redhat.com >> https://www.redhat.com/mailman/listinfo/sos-devel >> > The easiest way to get these reviewed and pulled in would be to fork the > project on github and create your pull requests that way. Unfortunately > I do not have admin priv's to the master repo or I'd take a look at it To be honest I find it easier to review patches via a mailing list - github is OK if you like doing things through a browser but it doesn't make it any easier if you're trying to work from a terminal. Regards, Bryn. From bmr at redhat.com Thu Nov 8 09:43:39 2012 From: bmr at redhat.com (Bryn M. Reeves) Date: Thu, 08 Nov 2012 09:43:39 +0000 Subject: [sos-devel] Gluster component update for sos In-Reply-To: <5099F417.80004@redhat.com> References: <5086D9EC.8070102@redhat.com> <5099F417.80004@redhat.com> Message-ID: <509B7ECB.8080101@redhat.com> On 07/11/12 05:39, Amar Tumballi wrote: > On 10/23/2012 11:24 PM, Amar Tumballi wrote: >> Hi all, >> >> I am attaching two patches which we are planing to have as part of sos >> for RHS2.0 (based on sos-2.2_17.el6_2). >> >> 0001-* was the first patch, done to perform 'statedump' only on one >> machine and to avoid 'gluster volume statedump'/'gluster volume status >> *' commands from the gluster component, with which we saw couple of issues. It might be better in future to use a randomly-generated temporary path as most other tools do. Generally it's good to avoid predictable temporary file names but this can be fixed later. I'd also like to lose the timeout thing but this needs changes in gluster - ideally we call a command, files are generated and collected. Even better if the gluster change includes the ability to specify an output path (that is within the existing sos report root) but again I'm fine with this temporarily since it only affects gluster users. >> We had couple of comments from Bryn on this patch, and the result is >> 0002-* patch which tried to remove the usage of /tmp/* directories. Also >> fixed an issue which Bryn saw on RHEL machines (without glusterd binary). Ack to these changes, thanks for getting this done. Regards, Bryn.