ygk_12345 | hi all | 08:02 |
---|---|---|
ygk_12345 | my rally task is failing with a pythin error | 08:02 |
ygk_12345 | *python | 08:02 |
ygk_12345 | 2021-12-28 13:26:56.485 71834 ERROR rally.task.engine TypeError: 'NoneType' object is not callable | 08:03 |
ygk_12345 | I am using python 3.9 | 08:03 |
holtgrewe | ygk_12345: can you put a full traceback to https://paste.openstack.org? | 08:10 |
ygk_12345_ | https://paste.opendev.org/show/811862/ | 08:15 |
ygk_12345_ | thats the rally task error | 08:15 |
ygk_12345_ | can someone check it ? | 08:15 |
ygk_12345_ | it is traced to multiprocessing module in python according to some links on the internet | 08:15 |
holtgrewe | ygk_12345: I'm only yet another user. The rally repo indicates they have a chat on Gitter. https://gitter.im/rally-dev/Lobby | 08:16 |
ygk_12345_ | holtgrewe: is it a known issue ? I am unable to open the gitter link due to filter here | 08:17 |
holtgrewe | ygk_12345_: I don't use rally. | 08:18 |
ygk_12345_ | @anyone here using rally ? | 08:18 |
holtgrewe | ygk_12345_: people might be on winter break for a few more days | 08:19 |
ygk_12345_ | hmm | 08:20 |
frickler | rally isn't tested with py3.9 yet, can you try with py3.8 instead? | 08:21 |
ygk_12345_ | frickler: it failed with 3.8 as well | 08:33 |
frickler | ygk_12345_: with the same error? can you share your job configuration? | 08:40 |
ygk_12345_ | frickler: yes same error | 08:41 |
ygk_12345_ | frickler: https://paste.opendev.org/show/811862/ | 08:43 |
frickler | ygk_12345_: that error is still showing py3.9. | 09:08 |
ygk_12345_ | frickler: yes, even though I run in virtualenv and set it to 3.8,, it is still showing 3.9. why ? | 09:09 |
ygk_12345_ | frickler: ah yes even with 3.8 the same issue | 09:10 |
gueswhat | guys, anyone running magnum in xena ? | 09:20 |
ygk_12345 | i am facing this error in rally | 13:12 |
ygk_12345 | https://paste.opendev.org/show/811862/ | 13:12 |
hgy_ | Helloļ¼could I ask a question? | 13:23 |
hgy_ | In CentOS8, Do I need replace NetworkManager with network-scripts? | 13:23 |
hgy_ | I saw that someone said NetowrkManager will have some confilt with openstack network | 13:23 |
AndreyKurilin[m] | frickler: regarding docs.o.o : I do not want to kill data there since I'm hoping to finish https://review.opendev.org/c/openstack/rally-openstack/+/761147 someday and restore docs.o.o as a part of that change . Unfortunately, do not have much time for that task recently :( | 13:44 |
AndreyKurilin[m] | frickler: as for an issue asked by ygk_12345 - it looks like some mac os related issue. I do not have an ability to debug that stuff :( | 13:52 |
ygk_12345 | AndreyKurilin[m]: strange a bit that macos python has an issue | 13:52 |
AndreyKurilin[m] | yeah :( | 13:53 |
gueswhat | is possible to use cinder for root disks in nova ? or do i need to have storage on each compute node for root disk and then cinder for additional volume ? | 22:26 |
jamesdenton | gueswhat you can boot entirely from a cinder volume is so configured | 23:16 |
jamesdenton | *if | 23:16 |
gueswhat | jamesdenton: how can i set it ? whats difference between root disk and ephemeral disk in flavours? | 23:18 |
gueswhat | magnum and nova consumes local disk on compute node, but i would like to set cinder for both as primary storage for os boot disk | 23:19 |
jamesdenton | hmm, i would probably equate to a C: and D: drive, but either way, both are local and ephemeral and do not survive instance deletion. | 23:44 |
gueswhat | i am confused, i dont know how to use cinder as storage for root and ephemeral disks... , ofc it makes a sense to use cinder as additional volume ( block disk ) for instance | 23:46 |
jamesdenton | This might help: https://docs.openstack.org/nova/latest/user/launch-instance-from-volume.html#create-volume-from-image-and-boot-instance | 23:47 |
gueswhat | use case: fedora coreos qcow2 image and flavour with root disk 32g and ephemeral 64g would create /sda ( root ) and /sdb ( ephemeral ) , both will be created on compute node and another volume ( sdc ) will be created in cinder, right ? | 23:48 |
gueswhat | jamesdenton: this is working only for instances, its not possible to override this in magnum | 23:50 |
gueswhat | maybe nfs mounted to compute node filepath and also nfs backend for cinder could make this working on single hdd | 23:53 |
jamesdenton | that might be true, i know that ceph can be used for default store but not sure about cinder. | 23:54 |
gueswhat | ceph for compute node and for cinder ? | 23:56 |
gueswhat | i would like to see a example with ceph for compute node, cinder, swift, manila, glance | 23:57 |
gueswhat | cuz, that would a big deal.. | 23:57 |
gueswhat | *be | 23:57 |
gueswhat | storage options is second most confusing thing about openstack | 23:58 |
jamesdenton | might have to stick around for someone else, i'm not super familiar myself | 23:59 |
Generated by irclog2html.py 2.17.3 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!