We have severity levels calculated for all legacy RHEL and Stronghold RHSA. We'd like to go back through all the released RHSA and change the current topic/synopsis lines to the new format which includes the severity information. This would require a little work from RHN to change the synopsis in their database, and a little work from us to change the corresponding one in the porkchop errata system. There are major business benefits in doing this; aside from customers being able to rank the legacy advisories it will help when studies compare our response times since we can point to the Red Hat severity level.
Created attachment 111303 [details] Proposed replacment synopsys lines
Lemme turn this back around on you... I jotted something down to see if we could start actually capturing this data as a part of rhn400... rather than just replacing the synopis, we could alter how we present errata to include this new field appropriately. Thoughts?
Yes, in the long term we'll present this field to RHN when pushing errata so that you can use it inside RHN for any number of neat data presentation ideas. However that requires changes to all the display mechanisms (logged in via rhn, /errata via rhn, up2date, rss feed) and so on. This is an interim measure that can be easily implemented immediately and will help counter the publicity that is expected at the start of next month when a new report is released surrounding our poor response times.
I've looked at webdev but there is a mistake on a few of these issues due to csv parsing... the ones which have ,'s in the data... for example RHSA-2002:248,Important: apache, mod_ssl, php security update for Stronghold you list this as "Important: apache" and drop everything after the second comma
Created attachment 111566 [details] quote the advisory names for csv processing
Mark, I've asked jslagle to revert the webdev change so we can setup a satellite pointing @ webdev and ensure that applying this change doesn't result in unexpected behavior (duplicate errata mails, etc). If it works ok, we'll apply the change, to get it out there in the near term, and revert it once we get an additional column to represent the data on the rhnErrata table.
updating estimates
If you get close to performing this update please let me know as I will check for a final time the replacement lines before you commit them.
Mark, do you want to provide me with a new csv, or just review the changes once I've applied them on dev? I plan on putting them on dev this afternoon. It will go to production on 7/19.
I can supply an updated csv tommorrow morning; can you wait for that?
Yes, that's no problem.
Created attachment 116402 [details] Replacent names
Okay, so I double checked the file this morning and modified it -- we decided not to change the synopsis of kernel issues that were part of an Update release for example. Attached (see comment 13)
The synopsis have been update on dev. Note: The following errata in the csv don't seem to exist in RHN: --No rows found for: -- RHSA-2002:040 Critical: php security update for Stronghold --No rows found for: -- RHSA-2002:045 Important: mod_ssl security update for Stronghold --No rows found for: -- RHSA-2004:297 Low: xchat security update Testplan ======== 1. Use the advanced search for Errata and search for some of the errata in the attached csv by advisory. 2. Verify that the synopsis for the errata are set to what is in the csv file. Keep in mind you may not be able to see all of these errata in a search, b/c you can only search for errata that are in a channel to which you have access.
I've tested this by looking at specific issues at random on rhn.webdev, and then looking at the list of all RHEL3, RHEL4, Stronghold, RHEL2.1 advisories at https://rhn.webdev.redhat.com/errata/ making sure that all issues were correctly marked. This found one missing entry: RHSA-2005:104,"Moderate: mod_python security update" RHSA-2004:297 can be ignored, it is a duplicate of 2004:585 RHSA-2002:040 and RHSA-2002:045 can both be ignored, they are "Stronghold Cross Platform" packages.
All the updates for the errata in the attachment are now on qa. The additional update in comment #16 is on_dev and will be pushed to qa during our next qa push, scheduled for 7/14.
On QA, looked at a random sampling at https://rhn.webqa.redhat.com/errata/ and all looked good.
looks good to me, prod ready