From David.Lee at ecmwf.int Mon Apr 23 10:30:56 2012 From: David.Lee at ecmwf.int (David Lee) Date: Mon, 23 Apr 2012 11:30:56 +0100 Subject: [rhn-users] Updating: RHel 5.x baseline and tailored updates or intermediate baselines Message-ID: <4F952F60.6040107@ecmwf.int> Preliminary question: is this list still actually active? The main question... We need to decide a philosophy of RHEL installation and updating to achieve stability and uniformity across a group of machines. The main question is about the baseline we should use for system deployment. (In other words, to be able to say "This is what we roll out and maintain consistently across the machines.") The two main options for establishing our baseline seem to be: 1. Pick some arbitrary, but fixed, point in time to generate a frozen local repo based on the RHN repo at that time; 2. Use an official RHEL 5.x release. In favour of the first (our point-in-time, from RHN) is that the baseline will (at the point of freezing) be reasonably up-to-date; i.e. notional minimisation of both bugs and integration instability. In favour of the second (official 5.x) is that such releases will have undergone more thorough integration testing. (Put the other way: arbitrary point-in-time snapshots have a slightly higher risk of creeping instabilities, as bugfixes etc. have the potential to introduce new problems, including unforeseen interactions with other areas.) Whichever we choose, we'll need to supplement this with a means to introduce (again from RHN) targeted, specific bugfixes for specific problems (e.g. security fixes) but I believe that is more or less independent of where we make the baseline for our systems. So advice (or pointers to papers offering advice) on where to place the baseline (RHN or ISO) would be welcome. If there is a more appropriate forum for this question, let me know. -- : David Lee : ECMWF (Data Handling System) : Shinfield Park : Reading RG2 9AX : Berkshire : : tel: +44-118-9499 362 : email: david.lee at ecmwf.int From Steven.Jones at vuw.ac.nz Mon Apr 23 20:26:37 2012 From: Steven.Jones at vuw.ac.nz (Steven Jones) Date: Mon, 23 Apr 2012 20:26:37 +0000 Subject: [rhn-users] Updating: RHel 5.x baseline and tailored updates or intermediate baselines In-Reply-To: <4F952F60.6040107@ecmwf.int> References: <4F952F60.6040107@ecmwf.int> Message-ID: <833D8E48405E064EBC54C84EC6B36E404CC82625@STAWINCOX10MBX1.staff.vuw.ac.nz> Hi, Why do you feel the need to fix a baseline? If you are going to do that then I would assume satellite will be the easiest way to manage this....but of course that costs and its not that great.........but it probably allows you to do what you want....ie have rolling baselines... So we have the base RH channel, then a testing channel and a prod channel....testing is 2 to 4 weeks behind RH base and prod about the same behind testing....but I doubt its of huge value as its a lot of work to maintain this.....on the other side of the coin however RH's quality isnt 100% perfect....ie they have minor screw ups.....so way up the time to have such a regime v the time to fix major issues...... regards Steven Jones Technical Specialist - Linux RHCE Victoria University, Wellington, NZ 0064 4 463 6272 ________________________________________ From: rhn-users-bounces at redhat.com [rhn-users-bounces at redhat.com] on behalf of David Lee [David.Lee at ecmwf.int] Sent: Monday, 23 April 2012 10:30 p.m. To: rhn-users at redhat.com Subject: [rhn-users] Updating: RHel 5.x baseline and tailored updates or intermediate baselines Preliminary question: is this list still actually active? The main question... We need to decide a philosophy of RHEL installation and updating to achieve stability and uniformity across a group of machines. The main question is about the baseline we should use for system deployment. (In other words, to be able to say "This is what we roll out and maintain consistently across the machines.") The two main options for establishing our baseline seem to be: 1. Pick some arbitrary, but fixed, point in time to generate a frozen local repo based on the RHN repo at that time; 2. Use an official RHEL 5.x release. In favour of the first (our point-in-time, from RHN) is that the baseline will (at the point of freezing) be reasonably up-to-date; i.e. notional minimisation of both bugs and integration instability. In favour of the second (official 5.x) is that such releases will have undergone more thorough integration testing. (Put the other way: arbitrary point-in-time snapshots have a slightly higher risk of creeping instabilities, as bugfixes etc. have the potential to introduce new problems, including unforeseen interactions with other areas.) Whichever we choose, we'll need to supplement this with a means to introduce (again from RHN) targeted, specific bugfixes for specific problems (e.g. security fixes) but I believe that is more or less independent of where we make the baseline for our systems. So advice (or pointers to papers offering advice) on where to place the baseline (RHN or ISO) would be welcome. If there is a more appropriate forum for this question, let me know. -- : David Lee : ECMWF (Data Handling System) : Shinfield Park : Reading RG2 9AX : Berkshire : : tel: +44-118-9499 362 : email: david.lee at ecmwf.int _______________________________________________ rhn-users mailing list rhn-users at redhat.com https://www.redhat.com/mailman/listinfo/rhn-users From Steven.Jones at vuw.ac.nz Mon Apr 23 20:52:04 2012 From: Steven.Jones at vuw.ac.nz (Steven Jones) Date: Mon, 23 Apr 2012 20:52:04 +0000 Subject: [rhn-users] Updating: RHel 5.x baseline and tailored updates or intermediate baselines In-Reply-To: <833D8E48405E064EBC54C84EC6B36E404CC82625@STAWINCOX10MBX1.staff.vuw.ac.nz> References: <4F952F60.6040107@ecmwf.int>, <833D8E48405E064EBC54C84EC6B36E404CC82625@STAWINCOX10MBX1.staff.vuw.ac.nz> Message-ID: <833D8E48405E064EBC54C84EC6B36E404CC82659@STAWINCOX10MBX1.staff.vuw.ac.nz> Sorry minor issues....in 7 years Ive never had a serious issue with RH's packaging... regards Steven Jones Technical Specialist - Linux RHCE Victoria University, Wellington, NZ 0064 4 463 6272 ________________________________________ From: rhn-users-bounces at redhat.com [rhn-users-bounces at redhat.com] on behalf of Steven Jones [Steven.Jones at vuw.ac.nz] Sent: Tuesday, 24 April 2012 8:26 a.m. To: Discussions about Red Hat Network (rhn.redhat.com) Subject: Re: [rhn-users] Updating: RHel 5.x baseline and tailored updates or intermediate baselines Hi, Why do you feel the need to fix a baseline? If you are going to do that then I would assume satellite will be the easiest way to manage this....but of course that costs and its not that great.........but it probably allows you to do what you want....ie have rolling baselines... So we have the base RH channel, then a testing channel and a prod channel....testing is 2 to 4 weeks behind RH base and prod about the same behind testing....but I doubt its of huge value as its a lot of work to maintain this.....on the other side of the coin however RH's quality isnt 100% perfect....ie they have minor screw ups.....so way up the time to have such a regime v the time to fix major issues...... regards Steven Jones Technical Specialist - Linux RHCE Victoria University, Wellington, NZ 0064 4 463 6272 ________________________________________ From: rhn-users-bounces at redhat.com [rhn-users-bounces at redhat.com] on behalf of David Lee [David.Lee at ecmwf.int] Sent: Monday, 23 April 2012 10:30 p.m. To: rhn-users at redhat.com Subject: [rhn-users] Updating: RHel 5.x baseline and tailored updates or intermediate baselines Preliminary question: is this list still actually active? The main question... We need to decide a philosophy of RHEL installation and updating to achieve stability and uniformity across a group of machines. The main question is about the baseline we should use for system deployment. (In other words, to be able to say "This is what we roll out and maintain consistently across the machines.") The two main options for establishing our baseline seem to be: 1. Pick some arbitrary, but fixed, point in time to generate a frozen local repo based on the RHN repo at that time; 2. Use an official RHEL 5.x release. In favour of the first (our point-in-time, from RHN) is that the baseline will (at the point of freezing) be reasonably up-to-date; i.e. notional minimisation of both bugs and integration instability. In favour of the second (official 5.x) is that such releases will have undergone more thorough integration testing. (Put the other way: arbitrary point-in-time snapshots have a slightly higher risk of creeping instabilities, as bugfixes etc. have the potential to introduce new problems, including unforeseen interactions with other areas.) Whichever we choose, we'll need to supplement this with a means to introduce (again from RHN) targeted, specific bugfixes for specific problems (e.g. security fixes) but I believe that is more or less independent of where we make the baseline for our systems. So advice (or pointers to papers offering advice) on where to place the baseline (RHN or ISO) would be welcome. If there is a more appropriate forum for this question, let me know. -- : David Lee : ECMWF (Data Handling System) : Shinfield Park : Reading RG2 9AX : Berkshire : : tel: +44-118-9499 362 : email: david.lee at ecmwf.int _______________________________________________ rhn-users mailing list rhn-users at redhat.com https://www.redhat.com/mailman/listinfo/rhn-users _______________________________________________ rhn-users mailing list rhn-users at redhat.com https://www.redhat.com/mailman/listinfo/rhn-users