Yedidyah Bar David has posted comments on this change. Change subject: packaging: setup: support compressed backup/restore ......................................................................
Patch Set 3: (1 comment) http://gerrit.ovirt.org/#/c/26073/3/packaging/setup/ovirt_engine_setup/database.py File packaging/setup/ovirt_engine_setup/database.py: Line 701: stdin=dump_proc.stdout, Line 702: stdout=output_file, Line 703: stderr=subprocess.PIPE, Line 704: close_fds=True, Line 705: ) > yes test1234 never ends so entire process group will not end, so it is expe > yes test1234 never ends so entire process group will not end, so it is > expected behaviour. Definitely unexpected by me. I expect the behavior of sh (or the kernel, for that matter) - that when the read process closes the pipe, the writer gets an error EPIPE. But python seems to not create real (kernel, static-sized buffer (actually controllable by fcntl since 2.6.35)) pipes with subprocess.PIPE but create a memory buffer which can grow large without limitation. But I am not arguing - if you simply claim that it's up to the user to make sure that no process in the pipe outputs in an endless loop, that's just fine with me. It's just need to be clearly documented because it's not the normal pipe behavior. Line 706: dump_stdout, dump_stderr = dump_proc.communicate() Line 707: compr_stdout, compr_stderr = compr_proc.communicate() Line 708: dump_rc = dump_proc.returncode Line 709: compr_rc = compr_proc.returncode -- To view, visit http://gerrit.ovirt.org/26073 To unsubscribe, visit http://gerrit.ovirt.org/settings Gerrit-MessageType: comment Gerrit-Change-Id: I684f025c9c1d29c701068ba967be5a2115c7a6fb Gerrit-PatchSet: 3 Gerrit-Project: ovirt-engine Gerrit-Branch: master Gerrit-Owner: Yedidyah Bar David <d...@redhat.com> Gerrit-Reviewer: Alon Bar-Lev <alo...@redhat.com> Gerrit-Reviewer: Sandro Bonazzola <sbona...@redhat.com> Gerrit-Reviewer: Yedidyah Bar David <d...@redhat.com> Gerrit-Reviewer: automat...@ovirt.org Gerrit-HasComments: Yes _______________________________________________ Engine-patches mailing list Engine-patches@ovirt.org http://lists.ovirt.org/mailman/listinfo/engine-patches
