yahoo-eng-team team mailing list archive
-
yahoo-eng-team team
-
Mailing list archive
-
Message #69556
[Bug 1735663] [NEW] privsep.daemon takes 9 seconds during taking volume attaching operation
Public bug reported:
When we did volume attaching operation (cinder/ceph as backend), cinder
took 5 seconds to finish its job, but nova took another 9s for running
privsep helper before calling into cinder, cause the total amount time
for attaching operation becomes ~15 seconds which seems too long for us.
Here is the log
2017-12-01 03:46:23.028 7 INFO nova.compute.manager [req-0e8c7f34-c0bd-4911-9022-a8747ebc03b8 a942d72fde65421f9d1897f1c23e6c34 289dbde7038f454396fde8b136414701 - default default] [instance: 04b47cd2-600b-40c1-9c3b-43e71898d1b9] Attaching volume 15ed8c33-cc00-4445-8d02-43807abc85b8 to /dev/vdb
2017-12-01 03:46:27.264 7 INFO oslo.privsep.daemon [req-0e8c7f34-c0bd-4911-9022-a8747ebc03b8 a942d72fde65421f9d1897f1c23e6c34 289dbde7038f454396fde8b136414701 - default default] Running privsep helper: ['sudo', 'nova-rootwrap', '/etc/nova/rootwrap.conf', 'privsep-helper', '--config-file', '/usr/share/nova/nova-dist.conf', '--config-file', '/etc/nova/nova.conf', '--privsep_context', 'os_brick.privileged.default', '--privsep_sock_path', '/tmp/tmpZkKVsQ/privsep.sock']
2017-12-01 03:46:28.160 7 INFO oslo.privsep.daemon [req-0e8c7f34-c0bd-4911-9022-a8747ebc03b8 a942d72fde65421f9d1897f1c23e6c34 289dbde7038f454396fde8b136414701 - default default] Spawned new privsep daemon via rootwrap
2017-12-01 03:46:28.106 1345 INFO oslo.privsep.daemon [-] privsep daemon starting
2017-12-01 03:46:28.109 1345 INFO oslo.privsep.daemon [-] privsep process running with uid/gid: 0/0
2017-12-01 03:46:28.111 1345 INFO oslo.privsep.daemon [-] privsep process running with capabilities (eff/prm/inh): CAP_SYS_ADMIN/CAP_SYS_ADMIN/none
2017-12-01 03:46:28.112 1345 INFO oslo.privsep.daemon [-] privsep daemon running as pid 1345
2017-12-01 03:46:32.119 7 INFO nova.compute.resource_tracker [req-3b3a701f-d6a6-4511-acf9-35877317a2df - - - - -] Final resource view: name=host-192-168-11-5 phys_ram=65500MB used_ram=768MB phys_disk=822GB used_disk=10GB total_vcpus=40 used_vcpus=1 pci_stats=[]
Any idea here?
** Affects: nova
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Yahoo!
Engineering Team, which is subscribed to OpenStack Compute (nova).
https://bugs.launchpad.net/bugs/1735663
Title:
privsep.daemon takes 9 seconds during taking volume attaching
operation
Status in OpenStack Compute (nova):
New
Bug description:
When we did volume attaching operation (cinder/ceph as backend),
cinder took 5 seconds to finish its job, but nova took another 9s for
running privsep helper before calling into cinder, cause the total
amount time for attaching operation becomes ~15 seconds which seems
too long for us.
Here is the log
2017-12-01 03:46:23.028 7 INFO nova.compute.manager [req-0e8c7f34-c0bd-4911-9022-a8747ebc03b8 a942d72fde65421f9d1897f1c23e6c34 289dbde7038f454396fde8b136414701 - default default] [instance: 04b47cd2-600b-40c1-9c3b-43e71898d1b9] Attaching volume 15ed8c33-cc00-4445-8d02-43807abc85b8 to /dev/vdb
2017-12-01 03:46:27.264 7 INFO oslo.privsep.daemon [req-0e8c7f34-c0bd-4911-9022-a8747ebc03b8 a942d72fde65421f9d1897f1c23e6c34 289dbde7038f454396fde8b136414701 - default default] Running privsep helper: ['sudo', 'nova-rootwrap', '/etc/nova/rootwrap.conf', 'privsep-helper', '--config-file', '/usr/share/nova/nova-dist.conf', '--config-file', '/etc/nova/nova.conf', '--privsep_context', 'os_brick.privileged.default', '--privsep_sock_path', '/tmp/tmpZkKVsQ/privsep.sock']
2017-12-01 03:46:28.160 7 INFO oslo.privsep.daemon [req-0e8c7f34-c0bd-4911-9022-a8747ebc03b8 a942d72fde65421f9d1897f1c23e6c34 289dbde7038f454396fde8b136414701 - default default] Spawned new privsep daemon via rootwrap
2017-12-01 03:46:28.106 1345 INFO oslo.privsep.daemon [-] privsep daemon starting
2017-12-01 03:46:28.109 1345 INFO oslo.privsep.daemon [-] privsep process running with uid/gid: 0/0
2017-12-01 03:46:28.111 1345 INFO oslo.privsep.daemon [-] privsep process running with capabilities (eff/prm/inh): CAP_SYS_ADMIN/CAP_SYS_ADMIN/none
2017-12-01 03:46:28.112 1345 INFO oslo.privsep.daemon [-] privsep daemon running as pid 1345
2017-12-01 03:46:32.119 7 INFO nova.compute.resource_tracker [req-3b3a701f-d6a6-4511-acf9-35877317a2df - - - - -] Final resource view: name=host-192-168-11-5 phys_ram=65500MB used_ram=768MB phys_disk=822GB used_disk=10GB total_vcpus=40 used_vcpus=1 pci_stats=[]
Any idea here?
To manage notifications about this bug go to:
https://bugs.launchpad.net/nova/+bug/1735663/+subscriptions
Follow ups