| -@gerrit:opendev.org- OpenStack Proposal Bot proposed: [openstack/project-config] 976568: Normalize projects.yaml https://review.opendev.org/c/openstack/project-config/+/976568 | 02:35 | |
| -@gerrit:opendev.org- Sei Sano proposed: [opendev/irc-meetings] 976571: Update Masakari team meeting agenda URL and frequency https://review.opendev.org/c/opendev/irc-meetings/+/976571 | 03:12 | |
| -@gerrit:opendev.org- Zuul merged: [openstack/project-config] 976568: Normalize projects.yaml https://review.opendev.org/c/openstack/project-config/+/976568 | 07:17 | |
| @harbott.osism.tech:regio.chat | dib is showing a weird failure with py313, seems something has changed in the implementation of `sort` or `cmp_to_key`. but I also think the comparator is broken since it claims that both `/var < /boot` and `/boot < /var` https://zuul.opendev.org/t/openstack/build/2d2178c7d1f14804961a76f6f1589d0f | 09:09 |
|---|---|---|
| @iwienand:matrix.org | Indeed | 09:56 |
| ``` | ||
| import functools | ||
| def cmp_mount_order(this, other): | ||
| """Sort comparision function for mount-point sorting | ||
| See if ``this`` comes before ``other`` in mount-order list. In | ||
| words: if the other mount-point has us as it's parent, we come | ||
| before it (are less than it). e.g. ``/var < /var/log < | ||
| /var/log/foo`` | ||
| :param this: tuple of mount_point, node name | ||
| :param other: tuple of mount_point, node name | ||
| :returns int: cmp value | ||
| """ | ||
| # sort is only based on the mount_point. | ||
| this, _ = this | ||
| other, _ = other | ||
| if this == other: | ||
| return 0 | ||
| if other.startswith(this): | ||
| return -1 | ||
| else: | ||
| return 1 | ||
| partitions = [ | ||
| ('/var/log', 'fake_log'), | ||
| ('/boot', 'fake_boot'), | ||
| ('/', 'fake_name'), | ||
| ('/var', 'fake_name')] | ||
| partitions.sort(key=functools.cmp_to_key(cmp_mount_order)) | ||
| print(partitions) | ||
| [iwienand@fedora19 tmp]$ python3.14 ./test.py | ||
| [('/', 'fake_name'), ('/var/log', 'fake_log'), ('/boot', 'fake_boot'), ('/var', 'fake_name')] | ||
| [iwienand@fedora19 tmp]$ python3.10 ./test.py | ||
| [('/', 'fake_name'), ('/var', 'fake_name'), ('/var/log', 'fake_log'), ('/boot', 'fake_boot')] | ||
| ``` | ||
| @harbott.osism.tech:regio.chat | yes, and the py314 result is correct in a sense, since `/ < /var/log < /boot < /var` according to `cmp_mount_order()`. I wonder how to best solve this, though. likely need to mix in some alphabetic ordering? or have `cmp_mount_order('/boot', '/var')==0`? | 10:05 |
| @iwienand:matrix.org | yeah ... https://github.com/python/cpython/issues/138246 | 10:06 |
| @iwienand:matrix.org | i think all we care about is the relative ordering of `/var` and `/var/log` | 10:14 |
| @iwienand:matrix.org | i.e. `/var/log` must be mounted after `/var/` | 10:14 |
| @iwienand:matrix.org | I feel like `"/var" < "/var/log"` | 10:16 |
| -@gerrit:opendev.org- Ian Wienand proposed: [openstack/diskimage-builder] 976593: block_device: remove cmp_mount_order https://review.opendev.org/c/openstack/diskimage-builder/+/976593 | 10:28 | |
| @harbott.osism.tech:regio.chat | ah, that GH issue is exactly what's happening, yes, broken transitivity | 10:29 |
| not sure if I got `<` and `>` mixed up, but the brokenness persists either way | ||
| ``` | ||
| def test_cmp(this, other): | ||
| if cmp_mount_order(this, other) == 1: | ||
| print("%s > %s" % (this, other)) | ||
| if cmp_mount_order(this, other) == -1: | ||
| print("%s < %s" % (this, other)) | ||
| if cmp_mount_order(this, other) == 0: | ||
| print("%s = %s" % (this, other)) | ||
| boot = ('/boot', 'foo') | ||
| var = ('/var', 'foo') | ||
| var_log = ('/var/log', 'foo') | ||
| test_cmp(var_log, boot) | ||
| test_cmp(boot, var) | ||
| test_cmp(var, var_log) | ||
| $ python3 test.py | ||
| ('/var/log', 'foo') > ('/boot', 'foo') | ||
| ('/boot', 'foo') > ('/var', 'foo') | ||
| ('/var', 'foo') < ('/var/log', 'foo') | ||
| ``` | ||
| @iwienand:matrix.org | Jens Harbott: ^ ? it feels right ... | 10:29 |
| @harbott.osism.tech:regio.chat | ah, making things simpler instead of even more complicated is a great idea, yes | 10:31 |
| @iwienand:matrix.org | my concern is why did we not just do that from the start, what am i missing 🤣 but all that block device code is ... well at least it has testing, let's just say that | 10:32 |
| -@gerrit:opendev.org- Eugene Gvozdetsky proposed: [opendev/git-review] 976601: Stop falling back to hardcoded 'master' branch https://review.opendev.org/c/opendev/git-review/+/976601 | 11:41 | |
| -@gerrit:opendev.org- Stephen Finucane proposed: | 12:29 | |
| - [openstack/project-config] 976612: Onboard more xstatic repos (1/2) https://review.opendev.org/c/openstack/project-config/+/976612 | ||
| - [openstack/project-config] 976613: Sync xstatic package to gerritbot https://review.opendev.org/c/openstack/project-config/+/976613 | ||
| - [openstack/project-config] 976614: Onboard more xstatic repos (2/2) https://review.opendev.org/c/openstack/project-config/+/976614 | ||
| -@gerrit:opendev.org- Stephen Finucane proposed: [openstack/project-config] 976612: Onboard more xstatic repos (1/2) https://review.opendev.org/c/openstack/project-config/+/976612 | 13:01 | |
| -@gerrit:opendev.org- Stephen Finucane proposed: | 13:02 | |
| - [openstack/project-config] 976612: Onboard more xstatic repos (1/2) https://review.opendev.org/c/openstack/project-config/+/976612 | ||
| - [openstack/project-config] 976613: Sync xstatic package to gerritbot https://review.opendev.org/c/openstack/project-config/+/976613 | ||
| - [openstack/project-config] 976614: Onboard more xstatic repos (2/2) https://review.opendev.org/c/openstack/project-config/+/976614 | ||
| -@gerrit:opendev.org- Takashi Kajinami proposed on behalf of Stephen Finucane: [openstack/project-config] 976612: Onboard more xstatic repos (1/2) https://review.opendev.org/c/openstack/project-config/+/976612 | 13:54 | |
| -@gerrit:opendev.org- Stephen Finucane proposed: [openstack/project-config] 976613: Sync xstatic package to gerritbot https://review.opendev.org/c/openstack/project-config/+/976613 | 13:55 | |
| -@gerrit:opendev.org- Stephen Finucane proposed: [openstack/project-config] 976614: Onboard more xstatic repos (2/2) https://review.opendev.org/c/openstack/project-config/+/976614 | 13:55 | |
| -@gerrit:opendev.org- Takashi Kajinami proposed on behalf of Stephen Finucane: [openstack/project-config] 976614: Onboard more xstatic repos (2/2) https://review.opendev.org/c/openstack/project-config/+/976614 | 13:55 | |
| -@gerrit:opendev.org- Zuul merged on behalf of Sei Sano: [opendev/irc-meetings] 976571: Update Masakari team meeting agenda URL and frequency https://review.opendev.org/c/opendev/irc-meetings/+/976571 | 14:33 | |
| -@gerrit:opendev.org- Zuul merged on behalf of Amy Marrich: [opendev/irc-meetings] 976555: Update week to the first Tuesday of the Month https://review.opendev.org/c/opendev/irc-meetings/+/976555 | 14:35 | |
| @clarkb:matrix.org | infra-root thoughts on https://review.opendev.org/c/opendev/gerritbot/+/975331 and its parent? I think we should land the parent first and ensure that connectivity continues to work. And if so landing 975331 should be pretty safe. Also I think we should do gerritbot before statusbot | 17:36 |
| @clarkb:matrix.org | but I'm happy to do statusbot first if we think that is a lower impact if the ssl connectivity stuff doesn't work | 17:36 |
| @fungicide:matrix.org | i was about to ask whether you wanted to do one or both outstanding irc bots today | 17:36 |
| @clarkb:matrix.org | fungi: yes I think that would be great. If anything braeks a revert should be quick and easy and overall impact is low outside of limnoria | 17:37 |
| @clarkb:matrix.org | My availability continues to be hit and miss but probably good enough for keeping an eye on $ircbot | 17:37 |
| @fungicide:matrix.org | yeah, i expect to disappear around 21z for an early dinner out, but have bandwidth to watch deployments and restart/revert things between now and then | 17:38 |
| @fungicide:matrix.org | approved them just now | 17:40 |
| @clarkb:matrix.org | great looks like you approved them | 17:40 |
| -@gerrit:opendev.org- Zuul merged on behalf of Stephen Finucane: | 17:44 | |
| - [openstack/project-config] 976612: Onboard more xstatic repos (1/2) https://review.opendev.org/c/openstack/project-config/+/976612 | ||
| - [openstack/project-config] 976613: Sync xstatic package to gerritbot https://review.opendev.org/c/openstack/project-config/+/976613 | ||
| -@gerrit:opendev.org- Zuul merged on behalf of Clark Boylan: [opendev/gerritbot] 975344: Use irc bot connection factory for SSL https://review.opendev.org/c/opendev/gerritbot/+/975344 | 17:45 | |
| @clarkb:matrix.org | The hourly deployment should get that updated in about 20 minutes | 17:46 |
| @fungicide:matrix.org | quick turnaround | 17:49 |
| @fungicide:matrix.org | the container image update hasn't merged yet though | 17:50 |
| @clarkb:matrix.org | It should merge shortly according to zuul | 17:52 |
| @fungicide:matrix.org | yeah, should make it in time | 17:52 |
| -@gerrit:opendev.org- Zuul merged on behalf of Clark Boylan: [opendev/gerritbot] 975331: Update container image to python3.12 and trixie https://review.opendev.org/c/opendev/gerritbot/+/975331 | 17:53 | |
| @fungicide:matrix.org | yay! | 17:58 |
| @clarkb:matrix.org | hrm I notice that opendevreview dropped off irc just after the change merged so maybe somethingdid deploy it? | 17:59 |
| @clarkb:matrix.org | ya it is running a 15 minute old image and is up for 12 ish minutes. That would be the first not second change | 18:00 |
| @clarkb:matrix.org | it just updated to the second change | 18:01 |
| @clarkb:matrix.org | I don't see anything obviously wrong yet, But haven't found any new messages posted to irc either so still trying to find evidence of function that way | 18:02 |
| @clarkb:matrix.org | ok https://meetings.opendev.org/irclogs/%23openstack-ansible/%23openstack-ansible.2026-02-12.log.html#openstack-ansible.2026-02-12.log.html#t2026-02-12T17:59:08 this message was after the first restart for sure. Not sure about the second | 18:04 |
| @clarkb:matrix.org | The message at 18:04 here https://meetings.opendev.org/irclogs/%23openstack-manila/%23openstack-manila.2026-02-12.log was after the second restart. Now I just need to c onfirm it used tls | 18:05 |
| @clarkb:matrix.org | doesn't look like we log the connectivity the same way we do with ptgbot. I can see it identified with nickserv. fungi do you know if we can ask the server if the client is connceted via tls? | 18:07 |
| @fungicide:matrix.org | i don't see a way | 18:09 |
| @fungicide:matrix.org | i guess we could check the remote port number? | 18:09 |
| @fungicide:matrix.org | irc typically uses pre-negotiated tls, not starttls over a plaintext socket | 18:10 |
| @clarkb:matrix.org | the configured port number is 6697 and that is what the change checks to determine using the ssl context wrapper | 18:10 |
| @clarkb:matrix.org | but if I got something wrong about that (type mismatch or whatever) then it could skip right over that. Maybe we just tcpdump -a and see if there is plaintext or tls? | 18:10 |
| @fungicide:matrix.org | right, so it's using ssl | 18:11 |
| @fungicide:matrix.org | oh, sure that's an option too | 18:11 |
| @fungicide:matrix.org | or just see if ss lists it | 18:11 |
| @clarkb:matrix.org | let me see if I can figure that out | 18:11 |
| @clarkb:matrix.org | oh yup ss seems to be sufficient `sudo ss -np | grep gerritbot` shows the connection to oftc on port 6697 | 18:12 |
| @clarkb:matrix.org | so I think this look good | 18:12 |
| @fungicide:matrix.org | awesome | 18:12 |
| -@gerrit:opendev.org- Zuul merged on behalf of Stephen Finucane: [openstack/project-config] 976614: Onboard more xstatic repos (2/2) https://review.opendev.org/c/openstack/project-config/+/976614 | 18:13 | |
| @clarkb:matrix.org | fungi: https://review.opendev.org/c/opendev/statusbot/+/975324 and parent are the equivalent statusbot changes (the two changes should be almost identical) | 18:14 |
| @fungicide:matrix.org | yeah, going through them already | 18:15 |
| @fungicide:matrix.org | and both approved | 18:16 |
| @fungicide:matrix.org | we got one! /var/log/apache2/docs.opendev.org_access.log has a hit on the tripwire, user agent claims to be "Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:134.0) Gecko/20100101 Firefox/134.0" and whois claims the ip address belongs to a "cyper defense scanning | 18:24 |
| @fungicide:matrix.org | company in france | 18:24 |
| @fungicide:matrix.org | so probably not an llm training crawler | 18:24 |
| @clarkb:matrix.org | Still great to see that this maybe works | 18:24 |
| @clarkb:matrix.org | And it isn't likely to be a human | 18:24 |
| @fungicide:matrix.org | but almost certainly a crawler that lied about its user agent and ignored our robots.txt | 18:25 |
| @clarkb:matrix.org | ++ | 18:25 |
| @fungicide:matrix.org | https://www.onyphe.io/ is the source | 18:25 |
| @fungicide:matrix.org | not sure if they chose the url because of the code comment or because of being listed in robots.txt | 18:27 |
| @clarkb:matrix.org | The source of the request resolves backwards to that name? | 18:28 |
| @fungicide:matrix.org | a whois lookup for the ip address of the client returns a lengthy explanation of what they're doing and refers to the url for their company | 18:28 |
| @clarkb:matrix.org | Got it | 18:29 |
| @fungicide:matrix.org | "We are conducting Internet-scale network scanning to provide information for cyber defense purposes. We scan the full IPv4 address space and part of IPv6 address space. We are in no way targeting you specifically, you are just part of what is connected on the Internet. Our complete list of our IP ranges is available here: https://www.onyphe.io/ip-ranges.txt Opt-out by sending your IP ranges at: abuse at onyphe dot io" | 18:30 |
| @fungicide:matrix.org | obviously a bit off-base since simple port scanning wouldn't be trying arbitrary urls (including doing host headers) | 18:31 |
| -@gerrit:opendev.org- Zuul merged on behalf of Clark Boylan: | 18:32 | |
| - [opendev/statusbot] 975343: Use irc bot connection factory for SSL https://review.opendev.org/c/opendev/statusbot/+/975343 | ||
| - [opendev/statusbot] 975324: Update container images to trixie and python3.12 https://review.opendev.org/c/opendev/statusbot/+/975324 | ||
| @clarkb:matrix.org | fungi: I'm sure they downplay what they are doing to make it sounds better | 18:39 |
| @clarkb:matrix.org | statusbot updated one eavesdrop with the last hourly run | 19:10 |
| @clarkb:matrix.org | trying to verify it used tls now as well | 19:10 |
| @clarkb:matrix.org | ss reports it is connected via port 6697. Goign to look in the logs and see if there is any logging confirmation that tls was used (but don't expect that given gerritbot's logs) | 19:11 |
| @clarkb:matrix.org | statusbot has debug logs like ptgbot which shows `NOTICE opendevstatus :*** Connected securely via TLSv1.3 TLS_AES_256_GCM_SHA384-256` | 19:13 |
| @fungicide:matrix.org | seems clear-cut to me | 19:18 |
| -@gerrit:opendev.org- Clark Boylan proposed on behalf of Ian Wienand: [openstack/diskimage-builder] 976593: block_device: remove cmp_mount_order https://review.opendev.org/c/openstack/diskimage-builder/+/976593 | 19:39 | |
| @fungicide:matrix.org | gonna head out now to grab an early dinner, but will check back in an hour or so in case anything's come up in the meantime | 21:12 |
| @clarkb:matrix.org | the main remaining trixie updates are limnoria (irc meetbot) and gitea. There is also the question about what to do about gear and I guess the gitea-init image. I'm inclined to stop building both of those two rather than continue to update them | 21:21 |
| @clarkb:matrix.org | corvus: can you think of any reason that gear's contanier image is something people would be relying on if still using software that we built? I guess only if running zuulv3? | 21:56 |
| @clarkb:matrix.org | maybe it would even have to be an older zuul | 21:57 |
| @jim:acmegating.com | nope. i think we can stop building it. and if someone wants to start again, great. | 22:00 |
| @clarkb:matrix.org | cool I'll update the change i pushed to remove the builds instead of updating them | 22:07 |
| -@gerrit:opendev.org- Zuul merged on behalf of Ian Wienand: [openstack/diskimage-builder] 976593: block_device: remove cmp_mount_order https://review.opendev.org/c/openstack/diskimage-builder/+/976593 | 22:10 | |
| -@gerrit:opendev.org- Clark Boylan proposed: [opendev/gear] 975335: Drop container image builds https://review.opendev.org/c/opendev/gear/+/975335 | 22:12 | |
| @fungicide:matrix.org | looks like nothing too exciting going on after dinner, turns out | 22:18 |
| @clarkb:matrix.org | dinner was probably more exciting | 22:22 |
| -@gerrit:opendev.org- Clark Boylan proposed: [opendev/system-config] 976693: Drop jinja-init and gitea-init container images and jobs https://review.opendev.org/c/opendev/system-config/+/976693 | 22:29 | |
| @clarkb:matrix.org | And this is the cleanup of jinja-init and gitea-init | 22:29 |
| @clarkb:matrix.org | Both of 975335 and 976693 should be low risk as nothing is using those images | 22:41 |
Generated by irclog2html.py 4.0.0 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!