[03:20] Hello! [03:21] I have a question regarding launchpadlib: [03:21] https://launchpad.net/+apidoc/devel.html#bug-addAttachment [03:22] There is a "content_type" parameter that is not mandatory. What does launchpadlib do to determine the content type if none is provided? [07:44] pieq, if a value is not provided for the 'content_type' parameter, Launchpad tries to guess it using the 'guess_content_type' function from 'zope.contenttype' (https://zopecontenttype.readthedocs.io/en/latest/api.html#zope.contenttype.guess_content_type) [07:45] Afaics, Launchpadlib does not pass a value for 'content_type' when it is not provided and this is handled on the backend. [08:14] guruprasad: thanks! [08:51] pieq, you are welcome! [12:10] hello :), is there an ETA for lgw01 builders being back online? [12:19] ricotz: they're intentionally offline and retained only for emergencies [12:19] lcy02 does seem to need a bit of attention right now, but I'd prefer to only enable lgw01 if lcy02 is entirely dead [12:22] (lgw01 is an older hardware generation; the datacentre it's in isn't quite marked for decommissioning yet, but we're generally trying to move stuff out of it where practical) [12:25] cjwatson, hi, does lgw01 provide more RAM for a builder instance? [12:25] ricotz: no [12:25] lgw01 is *older* than lcy02 and less good in most ways, though as it happens the core count, RAM, and disk configurations are the same [12:26] I am building libreoffice on impish which occasionally fails with a linker error, likely an OOM [12:26] then lgw01 will not help you [12:26] but re-tries lately result in the same failure consistently :( [12:26] it would just be slower [12:26] I see, could be less concurrent jobs [12:27] we've talked about upping builder RAM allocations to 16GiB across the board, but haven't done that yet (and need to ensure that it isn't going to cause any problems) [12:28] I see, I was about to ask for some current specs [12:28] so there are not special builders, like for the autopkgtests which more RAM [12:31] ricotz: No; that's difficult to arrange since for latency reasons we do the reset of a builder before we know what's going to be dispatched to it [12:32] And we don't at present maintain a "large" pool or something. We may have to at some point, but that would be a chunk of complexity across the board so I want to resist that as long as possible [12:33] thank you for the insights, I hoping the mentioned build succeeds at some point === locutusofborg_ is now known as LocutusOfBorg === tolecnal_ is now known as tolecnal