seleniumserver: enable firefox 68.0.2esr
this is the firefox we are currently using on testnodes. It's useful to be able to run test with exactly same version to debug issues. /reviewed-on nexedi/slapos!646
-
Owner
@luke I copied the snapshots of this test to https://[2001:67c:1254:57:c6e7::a2bd]:9080/cqc-wZuwZqkeca/snapshots/ and as a zip file https://[2001:67c:1254:57:c6e7::a2bd]:9080/cqc-wZuwZqkeca/snapshots.zip because it's more than 1Go in total ...
-
mentioned in merge request slapos.core!145 (merged)
-
Owner
aah and when test fail during setUpClass we don't keep instance, because tear down is not called when setup failed. So we don't have anything left for slapos-master test
😞 -
Maintainer
@luke I copied the snapshots of this test to https://[2001:67c:1254:57:c6e7::a2bd]:9080/cqc-wZuwZqkeca/snapshots/ and as a zip file https://[2001:67c:1254:57:c6e7::a2bd]:9080/cqc-wZuwZqkeca/snapshots.zip because it's more than 1Go in total ...
Can we by default compress the snapshots?
aah and when test fail during setUpClass we don't keep instance, because tear down is not called when setup failed. So we don't have anything left for slapos-master test
😞 Oh, that's why I had no snapshots in my branch :(
So let's put the snapshot creation to
_cleanup
? -
Owner
Can we by default compress the snapshots?
Yes, but it's still 134Mo compressed and the plan is that in the end we can see these snapshots directly on http, so it's nice to keep individual files.
I think we don't need to keep so many logs from slapos node software and instance, this is what is big. That configuration is not good for the case of tests. I did not think before setting this configuration. They are same the default values ( http://supervisord.org/configuration.html?highlight=stdout_logfile_maxbytes ). For tests, I feel the last 1Mo is enough, but there might be other use cases where we want more history of logs.
How do you think about this ?
- adding an option like "keep logs" in standalone.py, if it's false, we set
stdout_logfile_maxbytes
to 1MB andstdout_logfile_backups
to 0 - in testcase.py, setting keep logs to True if
SLAPOS_TEST_VERBOSE
is true
We'll have to make the size of each snapshot smaller and maybe also remove duplicates, it's easy with https://pypi.org/project/fastdupes/
import fastdupes import os fastdupes.out.write = lambda *args, **kw: None # disable output groups = fastdupes.find_dupes(['/tmp/snapshots/snapshots/'], exact=True) for f, dups in groups.items(): dups.discard(f) for dup in dups: os.unlink(dup) os.link(f, dup)
$ du -hs snapshots/ 3.1G snapshots/ $ find snapshots/ | sort > before $ python dedup.py $ du -hs snapshots/ 103M snapshots/ $ diff before <(find snapshots/ | sort)
So let's put the snapshot creation to _cleanup?
I see you started slapos.core!150 (merged) for this
👍 - adding an option like "keep logs" in standalone.py, if it's false, we set
-
mentioned in merge request slapos.core!150 (merged)