Re: [Spacewalk-devel] Query question regarding system.listLatestUpgradeablePackages
Jeremy D Davis wrote: % Attached you will find the patch as per the documentation % https://fedorahosted.org/spacewalk/wiki/GitGuide . Sorry about all the % emails and not following documentation. ... % Hello Spacewalk Developers, % % After asking the last question I decided to go ahead and make a change % of my own to see if I can improve performance. I was able to improve % the query by at least 6 seconds faster. The old query was taking about % 6101 ms to finish where I my query takes about 161 ms. Much % improvement. Please let me know if there is anything I need to change. % If everything looks good please let me know how I can get this added % for the next release of Spacewalk (Prefer 2.1). Hello Jeremy, Thank you for your patch. I've reviewed it and completely agree with you. Moreover looking into this query I found out some more places which can be optimized so I did even more complex rewrite of it. And there was also very similar query for list on UpgradableList.do page. Both changes are in 2a11191e20de5d0fa751f61fd59e82d903242238 so you can check it. Regards, -- Michael Mráka Satellite Engineering, Red Hat ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] [PATCH] Refeactored cookie parsing code to better handle malformed cookies
Michael Dorman wrote: % We recently discovered an issue where requests for any pxt page would % generate 500 errors. Traced it back to this submitted bug: % https://bugzilla.redhat.com/show_bug.cgi?id=723372 % % Apache2::Cookie-fetch dies if there is a malformed cookie sent by the % browser. In our case, it was a wildcard cookie, set by a different % site with the same domain name as our Spacewalk installation. So this % is not due to a bad cookie that Spacewalk itself is setting, but % rather the inability to deal with a malformed cookie set elsewhere. Hello Michael, I've actually already commited your patch but I have to revert it because it causes internal server error e.g. on /help/about.pxt. Regards, -- Michael Mráka Satellite Engineering, Red Hat ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] [PATCH] Refeactored cookie parsing code to better handle malformed cookies
Fair enough. I will take a look and see if I can address that. If not, this is really not that critical since most people are unlikely to hit this problem, anyway. On 2/20/14 5:19 AM, Michael Mraka michael.mr...@redhat.com wrote: Michael Dorman wrote: % We recently discovered an issue where requests for any pxt page would % generate 500 errors. Traced it back to this submitted bug: % https://bugzilla.redhat.com/show_bug.cgi?id=723372 % % Apache2::Cookie-fetch dies if there is a malformed cookie sent by the % browser. In our case, it was a wildcard cookie, set by a different % site with the same domain name as our Spacewalk installation. So this % is not due to a bad cookie that Spacewalk itself is setting, but % rather the inability to deal with a malformed cookie set elsewhere. Hello Michael, I've actually already commited your patch but I have to revert it because it causes internal server error e.g. on /help/about.pxt. Regards, -- Michael Mráka Satellite Engineering, Red Hat ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] Query question regarding system.listLatestUpgradeablePackages
Hello Michael, Thank you for looking at this and improving it further. I have pulled the latest patches into my branch and rebuilt the packages for both java and schema. I then performed an update of these new packages and performed a schema update on the DB but it seems the new arch ID for the view didn't get created. Is there anything I need to do in order to get this created to allow the new queries to run? Thank you for your time and have a great day! Regards, Jeremy -Original Message- From: spacewalk-devel-boun...@redhat.com [mailto:spacewalk-devel-boun...@redhat.com] On Behalf Of Michael Mraka Sent: Thursday, February 20, 2014 4:05 AM To: spacewalk-devel@redhat.com Subject: Re: [Spacewalk-devel] Query question regarding system.listLatestUpgradeablePackages Jeremy D Davis wrote: % Attached you will find the patch as per the documentation % https://fedorahosted.org/spacewalk/wiki/GitGuide . Sorry about all the % emails and not following documentation. ... % Hello Spacewalk Developers, % % After asking the last question I decided to go ahead and make a change % of my own to see if I can improve performance. I was able to improve % the query by at least 6 seconds faster. The old query was taking about % 6101 ms to finish where I my query takes about 161 ms. Much % improvement. Please let me know if there is anything I need to change. % If everything looks good please let me know how I can get this added % for the next release of Spacewalk (Prefer 2.1). Hello Jeremy, Thank you for your patch. I've reviewed it and completely agree with you. Moreover looking into this query I found out some more places which can be optimized so I did even more complex rewrite of it. And there was also very similar query for list on UpgradableList.do page. Both changes are in 2a11191e20de5d0fa751f61fd59e82d903242238 so you can check it. Regards, -- Michael Mráka Satellite Engineering, Red Hat ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] Query question regarding system.listLatestUpgradeablePackages
Michael, I was able to get it working by manually changing the rhnServerOutdatedPackages.sql to include the arch in the select and then remove the table from Spacewalk and run the query. Once I got that fixed I was able to test the queries you updated. It seem in my environment system_upgradable_packages_list_no_errata_info pulled duplicate packages due to a child channel contains these packages but is not assigned to the server being checked. So I have created a couple patches attached to this email that should resolve the issues for me. I am still looking for a way that I can deploy this without having to remove a table. Regards, Jeremy -Original Message- From: spacewalk-devel-boun...@redhat.com [mailto:spacewalk-devel-boun...@redhat.com] On Behalf Of Jeremy D Davis Sent: Thursday, February 20, 2014 10:37 AM To: spacewalk-devel@redhat.com Subject: Re: [Spacewalk-devel] Query question regarding system.listLatestUpgradeablePackages Hello Michael, Thank you for looking at this and improving it further. I have pulled the latest patches into my branch and rebuilt the packages for both java and schema. I then performed an update of these new packages and performed a schema update on the DB but it seems the new arch ID for the view didn't get created. Is there anything I need to do in order to get this created to allow the new queries to run? Thank you for your time and have a great day! Regards, Jeremy -Original Message- From: spacewalk-devel-boun...@redhat.com [mailto:spacewalk-devel-boun...@redhat.com] On Behalf Of Michael Mraka Sent: Thursday, February 20, 2014 4:05 AM To: spacewalk-devel@redhat.com Subject: Re: [Spacewalk-devel] Query question regarding system.listLatestUpgradeablePackages Jeremy D Davis wrote: % Attached you will find the patch as per the documentation % https://fedorahosted.org/spacewalk/wiki/GitGuide . Sorry about all the % emails and not following documentation. ... % Hello Spacewalk Developers, % % After asking the last question I decided to go ahead and make a change % of my own to see if I can improve performance. I was able to improve % the query by at least 6 seconds faster. The old query was taking about % 6101 ms to finish where I my query takes about 161 ms. Much % improvement. Please let me know if there is anything I need to change. % If everything looks good please let me know how I can get this added % for the next release of Spacewalk (Prefer 2.1). Hello Jeremy, Thank you for your patch. I've reviewed it and completely agree with you. Moreover looking into this query I found out some more places which can be optimized so I did even more complex rewrite of it. And there was also very similar query for list on UpgradableList.do page. Both changes are in 2a11191e20de5d0fa751f61fd59e82d903242238 so you can check it. Regards, -- Michael Mráka Satellite Engineering, Red Hat ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel 0001-Changed-the-system_upgradable_package_list_no_errata.patch Description: 0001-Changed-the-system_upgradable_package_list_no_errata.patch ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel
Re: [Spacewalk-devel] [PATCH] Refeactored cookie parsing code to better handle malformed cookies
Ok, here's an updated patch. The $req-jar call results in undef if there are no cookies at all, which it was not checking for. Thanks, Mike On 2/20/14 7:48 AM, Michael Dorman mdor...@godaddy.com wrote: Fair enough. I will take a look and see if I can address that. If not, this is really not that critical since most people are unlikely to hit this problem, anyway. On 2/20/14 5:19 AM, Michael Mraka michael.mr...@redhat.com wrote: Michael Dorman wrote: % We recently discovered an issue where requests for any pxt page would % generate 500 errors. Traced it back to this submitted bug: % https://bugzilla.redhat.com/show_bug.cgi?id=723372 % % Apache2::Cookie-fetch dies if there is a malformed cookie sent by the % browser. In our case, it was a wildcard cookie, set by a different % site with the same domain name as our Spacewalk installation. So this % is not due to a bad cookie that Spacewalk itself is setting, but % rather the inability to deal with a malformed cookie set elsewhere. Hello Michael, I've actually already commited your patch but I have to revert it because it causes internal server error e.g. on /help/about.pxt. Regards, -- Michael Mráka Satellite Engineering, Red Hat ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel 0001-Updated-cookie-parsing-code-to-handle-malformed-cook.patch Description: 0001-Updated-cookie-parsing-code-to-handle-malformed-cook.patch ___ Spacewalk-devel mailing list Spacewalk-devel@redhat.com https://www.redhat.com/mailman/listinfo/spacewalk-devel