=== praneshp_ is now known as praneshp | ||
=== harlowja is now known as harlowja_away | ||
=== pquerna_ is now known as pquerna | ||
mhroncok | smoser: hi, got a minute? | 11:08 |
---|---|---|
smoser | mhroncok, whats up? | 13:20 |
mhroncok | smoser: I wan to talk with you about cloud-init and Python 3 | 13:21 |
mhroncok | smoser: cloud-init is the last thing in Fedora cloud stack, that doesn't support Python 3 | 13:21 |
mhroncok | smoser: and we would like to make it happen | 13:21 |
mhroncok | how can we help? | 13:22 |
smoser | its on a list, and it should have been done alrady. | 13:22 |
smoser | i'm open to patches for that. we've done some work in that direction (dropping boto and cheetah recently) | 13:23 |
mhroncok | there are several relevant issues opened, but it seems nothing is happening, so we we would be happy to help with patches - however we are not sure what should we focus on first | 13:23 |
mhroncok | BTW boto now supports Python 3 | 13:23 |
mhroncok | I've seen this: https://code.launchpad.net/~harlowja/cloud-init/changeable-templates/+merge/208994 | 13:24 |
mhroncok | so now it should only be about cloud-init itself? | 13:25 |
smoser | i didn't know about boto and python3, thats good. yeah, it should be only cloud-init itself. | 13:26 |
mhroncok | is there any branch for python 3 porting or stuff liek tyhat, or we should just start with trunk? | 13:27 |
mhroncok | s/liek tyhat/like that/ | 13:27 |
smoser | trunk. | 13:27 |
smoser | harlowja_away, may have taken some further look at it. | 13:28 |
mhroncok | smoser: ok, thanks for info | 13:28 |
smoser | i think it makes sense to get to python3 support in cloud-init. | 13:30 |
smoser | there is some discussion about doing a cloud-init 2.0, which would start with python3 as the starting point. | 13:31 |
smoser | and probably not bother with python2 support | 13:31 |
mhroncok | smoser: is there a way to send patches, but don not sign CCA? | 13:31 |
smoser | i'm sorry, there is not. | 13:33 |
smoser | contributors agreement are required i many projects. | 13:33 |
mhroncok | I'm not even sure I can sign that, will consult my boss :) | 13:34 |
jfontan | hi | 14:14 |
jfontan | I am doing some modifications to a cloud init datasource | 14:14 |
jfontan | but I am unable to get the error | 14:15 |
jfontan | I only get a warning saying "Getting data from <clas... | 14:15 |
jfontan | I only get a warning saying "Getting data from <clas...> failed | 14:15 |
jfontan | is there any way I can see the errors? | 14:16 |
smoser | jfontan, maybe /var/log/cloud-init.log has an error ? | 14:38 |
jfontan | nope, that is the only error | 14:45 |
jfontan | I've added a traceback print to may copy of cloud-init | 14:46 |
jfontan | in find_source | 14:46 |
jfontan | may=my | 14:46 |
smoser | jfontan, you should get something else along side that error. | 16:03 |
jfontan | it only said it failed | 16:06 |
jfontan | it was an error in the code | 16:06 |
jfontan | but the exception blod just printed that it could not get the data | 16:07 |
jfontan | block | 16:07 |
jfontan | here | 16:08 |
jfontan | http://bazaar.launchpad.net/~cloud-init-dev/cloud-init/trunk/view/head:/cloudinit/sources/__init__.py#L255 | 16:08 |
smoser | harmw, what do you need me to merge ? | 16:57 |
smoser | jfontan, yeah, but it should log a stack trace | 16:57 |
jfontan | it didn't for me | 16:57 |
smoser | can you pastebinit ? | 17:02 |
smoser | the whole log ? | 17:02 |
jfontan | I'll have to create the image again | 17:08 |
jfontan | as I don't have the error in the code anymore and it has the trace print in. I'll check if I have an old vm | 17:10 |
jfontan | smoser, https://gist.github.com/jfontan/bcf18f262cd06252ff1a#file-gistfile1-txt-L13 | 17:18 |
jfontan | it only says that it failed but no stacktrace | 17:19 |
jfontan | I'm using trunk | 17:19 |
smoser | so the error you see there is the one you added ? | 17:20 |
smoser | the stack trace | 17:20 |
harlowja | smoser did i hear python3 | 17:21 |
harlowja | lol | 17:21 |
jfontan | nope, I took out the strack trace printing code | 17:21 |
jfontan | the error is in OpenNebula Data source | 17:21 |
jfontan | and just tells it failed | 17:21 |
jfontan | no stacktrace | 17:22 |
smoser | hm.. that is odd. i dont know why. | 17:22 |
jfontan | there is an stacktrace for openstack but that's something I'm not working on | 17:22 |
smoser | yeah, and that is wrong. | 17:22 |
smoser | you're using turnk her e? | 17:22 |
jfontan | yes | 17:22 |
smoser | harlowja, you see that ^ . | 17:22 |
smoser | that should probably not be a fail | 17:22 |
harlowja | whats the trace? | 17:22 |
smoser | but rather a "no datasource here, move along" | 17:22 |
smoser | https://gist.githubusercontent.com/jfontan/bcf18f262cd06252ff1a/raw/960d80f5263ebbe909ab786ef1ece8bc5d04cda6/gistfile1.txt | 17:23 |
harlowja | smoser hmmm, will fix | 17:23 |
jfontan | it's a failure. I mistyped a variable | 17:23 |
smoser | and any idea why we wouldnt get a log | 17:23 |
smoser | of that failure there. | 17:23 |
smoser | for the 'DataSourceOpenNebula' line | 17:24 |
jfontan | it could be the code in data source | 17:24 |
jfontan | maybe it's catching all exceptions | 17:24 |
smoser | well, it is. | 17:24 |
jfontan | I'll take a look | 17:24 |
smoser | but then it calls | 17:25 |
smoser | util.logexc(LOG, "Getting data from %s failed", cls) | 17:25 |
jfontan | so that logexc should print the trace, isn't it? | 17:26 |
smoser | yeah. | 17:26 |
jfontan | I have to go | 17:30 |
jfontan | good luck | 17:30 |
harlowja | smoser https://code.launchpad.net/~harlowja/cloud-init/mini-os-fixes/+merge/233399 | 17:37 |
harlowja | mhroncok also https://code.launchpad.net/~harlowja/cloud-init/py2-3 for your py3 messages | 18:02 |
harlowja | thats the branch that imho will enable mixed py2/py3 mode | 18:02 |
harlowja | which will be required for older distros that aren't all py3 yet | 18:03 |
mhroncok | harlowja: great, thanks | 18:03 |
harlowja | i did some basic testing from what i remember with py3 and pretty sure i nailed most of the issues down | 18:03 |
mhroncok | BTW harlowja smoser, cannot I just give you my patches as Public domain? | 18:03 |
harlowja | mhroncok smoser probably knows better than i all that stuff, i never got into the legal stuff :-P | 18:04 |
smoser | mhroncok, i dont think so. | 18:04 |
mhroncok | ok, waiting to hear from our legal :D | 18:04 |
smoser | harlowja, i don think you fixed https://gist.githubusercontent.com/jfontan/bcf18f262cd06252ff1a/raw/960d80f5263ebbe909ab786ef1ece8bc5d04cda6/gistfile1.txt | 19:03 |
smoser | in mini-os | 19:03 |
smoser | mini-os-fixes | 19:03 |
harlowja | k, maybe i misunderstand ? | 19:04 |
smoser | so in that log | 19:04 |
smoser | hm.. | 19:04 |
smoser | oldd. | 19:04 |
harlowja | 'openstack.py[ERROR]: Failed reading mandatory path /tmp/tmpX2lt9o/openstack/latest/meta_data.json' ? | 19:04 |
harlowja | thats now demoted to debug | 19:04 |
smoser | i have to look at this more. i didn't hitnk he had a config drive there. | 19:04 |
smoser | oh i see. | 19:05 |
smoser | so that would be debug, but then there is / should be no ERROR | 19:05 |
harlowja | _fetch_available_versions() takes exactly 1 argument (2 given) also should be fixed | 19:05 |
smoser | agreed there. | 19:05 |
smoser | i thought what had happend there was that there was no config drive | 19:05 |
smoser | and that should result in the DSs' _get_data() returnning False | 19:06 |
smoser | with only debug messages | 19:06 |
harlowja | sure, i just think the differeniation between searching for get_data that will work and actually using a get_data that then fails isn't clear | 19:07 |
harlowja | thats probably part of it | 19:07 |
harlowja | maybe should be refactored to find_data (which then only outputs debug messages) and get_data (which actually outputs errors) | 19:07 |
harlowja | sepearting the non-failure find process from the actual problems get_data | 19:08 |
harlowja | there sorta munged together at this point | 19:08 |
smoser | well, _get_data returns True on "I've found" | 19:16 |
smoser | and False on "nope" | 19:17 |
harlowja | right, but say it had a finding method, this one would return True/False, then it had a read method that would read from the previous location found (if it was) and would log errors and such | 19:19 |
harlowja | the find method would be expected to never log errors and such | 19:19 |
harlowja | but the read method could (instead of what we currently do is make everything debug, since we are doing both find and read in the same method) | 19:20 |
smoser | i agree, it can be better. | 19:25 |
smoser | harlowja, can you fix that though ? | 19:28 |
harlowja | hehe, what i'm supposed to fix? | 19:28 |
harlowja | lol | 19:28 |
smoser | it needs to be debug only on "not found" | 19:29 |
smoser | and ideally not stack trace :) | 19:29 |
harlowja | right, https://code.launchpad.net/~harlowja/cloud-init/mini-os-fixes/+merge/233399 switches those to debug ? | 19:30 |
harlowja | with no stack-trace | 19:30 |
harlowja | but i might still be confused, lol | 19:31 |
harlowja | oh, i think i see one more | 19:31 |
harlowja | ok, think i got them (reuploaded) | 19:34 |
smoser | how did it do no stacktrace | 20:00 |
harlowja | except openstack.NonReadable: | 20:03 |
harlowja | return False | 20:03 |
harlowja | probably that | 20:03 |
harlowja | except openstack.NonReadable: | 20:03 |
harlowja | return False | 20:03 |
harlowja | except (openstack.BrokenMetadata, IOError) as e: | 20:03 |
harlowja | LOG.debug("Broken metadata address %s: %s", | 20:03 |
harlowja | self.metadata_address, e) | 20:03 |
harlowja | return False | 20:03 |
harlowja | shall that be changed? | 20:04 |
harlowja | or similar in 'DataSourceConfigDrive' | 20:04 |
harmw | smoser: https://code.launchpad.net/~harmw/cloud-init/freebsd-fixes | 20:15 |
harmw | https://code.launchpad.net/~harmw/cloud-init/freebsd-fixes/+merge/233247 | 20:15 |
smoser | harmw, tried to build the cirros with patched busybox | 20:23 |
smoser | but busubox didn't get patched | 20:23 |
smoser | not sure why | 20:23 |
harmw | sorry man, this time it's me beeing way to busy :) | 21:07 |
smoser | harmw, no worries. | 21:16 |
smoser | it seems like an easy fix | 21:16 |
smoser | http://paste.ubuntu.com/8253590/ | 21:16 |
harmw | hm, that makefile change... can't recall i needed that | 21:53 |
smoser | no | 22:11 |
smoser | you didn't | 22:11 |
smoser | that was debug | 22:11 |
=== gondoi is now known as zz_gondoi |
Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!