[pca] Solaris 9 transitioning to Extended Support
Solaris 9 will transition to Vintage Support at the end of October. This means that any Solaris 9 patches (or patch revisions) released from Nov 1st onwards will require Vintage support level. Details on Gerry Haskins' blog: http://blogs.oracle.com/patch/entry/solaris_9_transitioning_to_extended BTW - when that happened with Solaris 8, I heard that Vintage support is very expensive. Did anybody ever buy and use that? Martin.
Re: [pca] Setting up PCA patch server
Hi all, I've thought about this kind of problem some weeks ago and figured out that on this topic, what I wanted from PCA was simple: Exactly the same as SUNWtlp was doing for me: - Bundle generation with frozen recommended set of patches - HTML Reports - Live Upgrade Friendly Bundle (with patch_order file and such) This way, you can simply use luupgrade -s /var/tmp/bundle /var/tmp/bundle/patch_order Theses were the first three feature I needed, amongst other that SUNWtlp was offering. For those who never seen SUNWtlp in action, it was a complete perl-based patching framework used to work in very large enterprise environment counting hundreds of servers. The principle was simple, a cronjob with a big batch perl script was going to every server and picks up the weekly explorer, extract it into TLP directory and then create a ready-to-use bundle of patches with associated HTML report. It was great and sufficient until SUN decided to decommission it in favor of xVM ops center. I've started myself to build a bunch of scripts to fulfill the feature I listed above. I don't know if this could be of some interest for any of you, but if it is, I've got nothing against publishing the code of theses scripts and eventually maintain it. To be clearer, I'll add a current output of what my scripts are doing... Regards, $ ./cleanExplorers [-] Removing explorer.82a91f35.uxhabsyc021-2010.06.19.01.00.tar.gz as it is older than 60 days [-] Removing explorer.8557b4a2.proatriusr02-2010.03.20.02.00.tar.gz as it is older than 60 days $ cp /tmp/explorer.847c9adf.uxvenusc012-2011.10.08.01.00.tar.gz /pca/data/explorers/ $ ./extractExplorers [-] Extracting ./explorer.8323e753.accatriu2oi11-2011.10.08.01.00.tar.gz $ ./batchReport [-] Generating report for accatriu2oi11... [-] Generating report for accatriu2po11... [-] Generating report for accatriudn04... **SNIPPED** $ ./pcaHost uxorionc012 -l missings Using /pca/data/patchxrefdir/patchdiag.xref from Oct/18/11 Host: uxorionc012 (SunOS 5.10/Generic_142909-17/sparc/sun4u) List: missings (66/10608) Patch IR CR RSB Age Synopsis -- -- - -- --- --- --- 118666 27 34 RS- 2 JavaSE 5.0: update 32 patch (equivalent to JDK 5.0u32) **SNIPPED** + bins to generate bundles, pack everything up, etc... -- Thomas Gouverneur _ _ | |___ _ __ (_)_ __ | _| / __| '_ \| \ \/ / | |___\__ \ |_) | | |_|___/ .__/|_/_/\_\ Network |_| SPRL TVA: BE6836018011 T: +32 498 23 00 40 W: http://espix.net M: tho...@espix.net On Tue, 18 Oct 2011 11:30:20 -0500 Becktold, Lisa M. (LARC-B7)[Chugach Federal Solutions, Inc.] lisa.m.beckt...@nasa.gov wrote: Hi, We have started using pca to patch our servers.It's been working very well, and we'd like to refine our setup. Instead of each server accessing support.oracle.com, we'd like to set up an internal pca patch server. The patch clients should access this internal patch server, and NOT go outside our internal network to download patches. I'm a bit confused about what should reside on that internal patch server. Should I download the latest recommended patch cluster and associated patchdiag.xref file, and place it on the internal patch server? If the full recommended patch cluster is located on the patch server, then there the patch client doesn't need to go outside our internal network and access support.oracle.com, right? The patch server must either share the patch directory via NFS, or offer it via http or ftp, correct? Lisa Lisa Becktold System Administrator Chugach Federal Solutions, Inc. NASA Center for AeroSpace Information Contract Phone: 443 995 5805
Re: [pca] Setting up PCA patch server
Lisa, Others have already described various options you have. As for my 2 cent: Take a look at the PCA documentation, especially the parts about setting up a local patch server (in your situation, a PCA local caching proxy might suit best, and there are detailed step-by-step instructions in the docs) and about creating patch reports for remote machines. As seen in other follow-ups, many people use PCA just as one tool in their patching framework. Therefore I kept PCA as general as possible, and you'll have to find your own way that fits best to your surroundings. I'm having a hard time getting my head around the fact that pca is very system-dependent. It has to be. A system may have any set of packages and patches installed, PCA won't assume that any two systems are exactly the same. So it sees what's (already) there, and tells you what's missing. If you as an administrator know that two systems are equal, it's fine to download the patches only once on one fo them and re-use them later on the second system. Martin.
Re: [pca] Recommended patches and PCA
Hi Jonathan, Thanks for the background information and the confirmation that --minimal works for you. I decided to make it official now, and have added documentation to the development release of PCA now. Without gushing too much we have found PCA too useful to not use. It simplifies and standardises everything for us. You made my day! :) Martin.
Re: [pca] Solaris 9 transitioning to Extended Support
@my company we weighed our options. it made more sense for us to move to version 10 than pay for extra support for version maintenance. Ggr On Oct 19, 2011 8:11 AM, Chuck Floyd ch...@cmfloyd.com wrote:
Re: [pca] Solaris 9 transitioning to Extended Support
Yes, Worked with them several times. Was not really a good experience but he could be that they did not understand the T3 storage systems here in Belgium. Hospital was down for more then a day before i came and started the reinstallation of the whole system because they managed to screw it up completly. Regards Filip Francis On 10/19/11 15:35, Thomas Gouverneur wrote: Anyone ever heard of Osiatis company that looks like a resseler of support contract for outdated systems ? My customer think of moving to such support contract as it's cheaper than Oracle ones. Any return of experience ? Rgds, Thomas On Wed, 19 Oct 2011 08:16:27 -0400 Rajiv Gunjaopn.src.ro...@gmail.com wrote: @my company we weighed our options. it made more sense for us to move to version 10 than pay for extra support for version maintenance. Ggr On Oct 19, 2011 8:11 AM, Chuck Floydch...@cmfloyd.com wrote:
Re: [pca] Solaris 9 transitioning to Extended Support
Fred wrote: We did the same thing. I do miss how fast Solaris 8 would reboot on our old V100's - down and up in 90 seconds. With Solaris 10 it takes about 4 minutes. Worry not, Solaris 11's fast reboot can boot a host as quickly as a zone. Default behaviour for a reboot is to skip the POST. Fred V100's won't run Solaris 11, and they don't run POST on a reboot anway. POST on our V100's takes about 45 seconds...a bit longer if they have more memory. -- Jeff Wieland| Purdue University Network Systems Administrator |ITNS Data Networks Voice: (765)496-8234 |155 S. Grant Street FAX: (765)494-6620| West Lafayette, IN 47907-2115