14:02:26 #startmeeting Security Team Meeting - Agenda: https://fedoraproject.org/wiki/Security_Team_meetings 14:02:26 Meeting started Thu Aug 6 14:02:26 2015 UTC. The chair is Sparks. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:02:26 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:02:29 #meetingname Fedora Security Team 14:02:29 The meeting name has been set to 'fedora_security_team' 14:02:32 #topic Roll Call 14:02:34 * Sparks 14:02:37 * d-caf 14:03:09 * mhayden 14:04:04 * scorneli 14:04:48 * zoglesby 14:05:45 * Sparks updates meeting agenda 14:06:27 Okay, lets get started. 14:06:57 #info Participants are reminded to make liberal use of #info #link #help in order to make the minutes "more better" 14:07:05 #topic Outstanding BZ Tickets 14:07:13 #info Thursday's numbers: Critical 0 (0), Important 45 (0), Moderate 373 (+12), Low 148 (-1), Total 520, Trend +11 14:07:19 #info Current tickets owned: 96 (~18%) 14:07:25 #info Tickets closed: 355 (+3) 14:07:36 Anyone have anything regarding tickets? 14:08:02 i think d-caf's tickets moved a bit 14:08:20 I've had a few move, after pestering people more 14:08:32 I feel like I'm nagging my kids sometimes... 14:08:54 jwb is moving patches for nagios-plugins which is good 14:09:03 * mhayden wonders if folks have these bugzilla notices auto-filtered into a directory that gets ignored ;) 14:09:13 though I'm still unclear as to what the eventual resolution is going to be longterm for that package 14:09:39 Would be nice if bugzilla let us knwo when messages bounced to email 14:09:54 I do filter bugzilla quite a lot. 90% is garbage, hard to find the good 10% 14:10:23 Hello, 14:10:30 pjp: hi 14:10:36 yo! 14:10:53 I tend to go out of bugzilla and email people direct, but it still can involve a lot of nagging on both fronts 14:11:17 but, then my actions are outside the tracking of bugzilla.. 14:11:34 yes. i still think we need a change of policy there 14:11:55 If people are ignoring BZ then they aren't being maintainers since they are likely ignoring other problems. 14:12:07 On the numbers info, would it be possible to break out the Important and maybe moderate with the precent owned? Vs just the overallowned percentage? 14:12:19 d-caf: Yes 14:12:19 Sparks: agree! 14:12:52 * Sparks to update the FST_numbers script to show the %age of tickets owned for each category. 14:13:20 i wonder if we can automate some notifications based on that script 14:13:26 #action Sparks to update the FST_numbers script to show the %age of tickets owned for each category. 14:13:41 mhayden: ? 14:13:48 I think there are still like 8-12 unowned important tickets 14:13:54 I picked up a few more this week 14:14:13 well i wonder if we could run a script monthly to roll through BZ and email maintainers about tickets w/CVE's that haven't moved in > 30 days 14:14:29 it would help if they have their BZ notifications filtered off somewhere 14:14:34 and we'd get bounces 14:14:39 mhayden: I have one script for that, 14:14:49 * jsmith shows up late, sorry 14:14:56 mhayden: it pings people after > 90 days of inactivity 14:15:24 jsmith: Welcome 14:15:24 * pjp would commit it to the fst security.git repo today 14:15:28 jsmith: Hi, 14:15:40 pjp: that's awesome! 14:16:02 pjp: You currently run that against your own tagged tickets or all? 14:16:08 Sparks: is the FST_numbers script available some place? 14:16:27 pjp: It's in the fedora-security-team repo 14:16:36 d-caf: all security bugs which haven't seen activity for > 90 days 14:16:53 pjp: And could probably be written much better but, instead, is just cobbled together bash 14:16:55 I started with that script as part of automating non-responsive maintainer policy 14:17:18 Sparks: Okay, I'll take a look 14:17:29 Ok, so if the only activity was FST members commenting, then it would run against those (so not ping the maintaners) 14:17:43 would/wouldn't 14:18:39 d-caf: Yeah that's where it gets a little tricky, when I ran it first time, some maintainers responded with comments about why bugs have not moved further, 14:19:13 d-caf: parsing and interpreting human comments is tricky, to script any further actions on that 14:19:50 * pjp looks for a sample bug 14:20:27 pjp: I would probably just set it to ignore the fst_owner=USER tagged comments/actions 14:20:56 d-caf: ? 14:21:27 * Sparks wonders if we should subscribe a mailing list to all security bugs so we can actually see comments like "I'm not going to fix this because your process sucks" so our automated tools won't count that as activity. 14:21:46 take the username mentioned in fst_owner and exclude comments from that user so that it only counts others comments as "refreshing" the ticket 14:22:04 d-caf: ex -> https://bugzilla.redhat.com/show_bug.cgi?id=1132002#c2 14:22:25 d-caf: #c2 was added by a script 14:23:10 d-caf: no-no, these comments were from bug owners 14:23:40 that's fine, if bug-ownders are commenting it means they know about it (but may just be dragging their feet) 14:23:54 that's a whole nother problem 14:24:14 d-caf: yes, 14:24:54 d-caf: I started with an idea that we ping them 3 times in a row, and if there is still no response, we open a non-responsive policy bug and automatically send an email to the -devel list 14:24:55 d-caf: Or ignoring the problem. 14:25:08 pjp: +1 14:25:24 if the script can email the assigne if no comments have been placed in the ticekt (exluding the fst_ownder comments) outside bugzilla 14:25:36 d-caf: But when they respond with some comments, that's tricky to parse 14:26:04 d-caf: Sure, email could be sent 14:26:15 Well if they responde then again we no they know about it, then it's about "pushing" them to fix it 14:26:25 respond then again we know... 14:26:29 ugh my typing sorry 14:26:37 :) 14:27:26 Another problem was, I got flooded with Bugzilla emails with the comments script added and what the bug owner responded 14:27:29 Things i would like to workaround: people ignoring bugzilla emails, peole with broken email addresses 14:28:00 Yep, those can be handled neatly, as there is not comment from them 14:28:11 Yeah, even with out your script i can get flooded with bugzilla emails 14:28:18 :) 14:35:24 #topic Open floor discussion/questions/comments 14:35:44 Anyone have anything else they'd like to discuss. 14:35:45 ? 14:36:14 i've considered trying to adapt the CIS benchmarks for fedora and build an image and/or an ansible playbook 14:36:23 Did we ever get that security@fedora email thing figured out? 14:36:23 not sure if that excites anyone else 14:37:26 d-caf: Yes, we now have control over the list and address, and I'm to some publicity work for it 14:37:37 s/to/to do some... 14:37:53 pjp: cool, thanks 14:37:58 d-caf: we need to let people know that such address exists and we are actively monitoring it 14:39:00 would want it to be easily found via google if you search for "fedora security" 14:39:48 Yes, true 14:41:34 mhayden: i haven't looked into those much before, will need to add it to my reading list 14:42:01 d-caf: CentOS/RHEL 6 is done --> https://github.com/major/cis-rhel-ansible 14:42:14 planning to expand that to RHEL 7 soon, but figured i might add logic for Fedora too 14:44:05 Cool work, I'm familiar with a couple other tools like this (like NIST) but hadn't really looked into CIS 14:47:36 * pjp checks 14:48:50 Heh...WAIT! DANGER! :) 14:52:16 I think we are reaching a closing time, 14:52:26 could we make a badge for folks who close CVE-related tickets? 14:52:30 yes I have nothing else 14:52:41 mhayden: There is a ticket open for that 14:52:51 oh that'd be fantastic 14:52:59 * mhayden has nothing else for the meetin' 14:53:15 mhayden: -> https://fedorahosted.org/fedora-badges/ticket/373 14:53:40 There is one for the logo too, which is in progress 14:54:20 mhayden: -> https://fedorahosted.org/fedora-badges/ticket/262 14:54:56 Please add comments to those tickets with any inputs if you have. 14:55:47 So that's for people who found and reported a CVE, not closed 14:56:21 d-caf: Yes, other one is for folks who contribute to FST activities, which could involve closing a CVE bug 14:56:40 If we want a separate one, then we'll need a ticket for it 14:57:01 * Sparks notes 4 minutes remain 14:57:06 * Sparks notes 3 minutes remain 15:00:30 Okay, we'll continue this discussion on the mailing list. Thanks everyone~! 15:00:33 #endmeeting