You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
By PR #5490 was introduced the degradation of ptf tests runtime and logic.
runtime:
Before the start of each ptf test, executed the checker for MACsec for each interface.
This step takes ~1:20 minutes.
For example the test ecmp/test_fgnhg.py:
Previously runtime of the test was ~24 minutes.
Now, the test terminated in our regression with a timeout of 1h. Where more than 30 times called the ptf test. It means that more than 30*1:20 =40 minutes from 1 hour we lost on MAcsec checker.
20:51:01 ptf.create_macsec_info L0043 INFO | MACsec isn't enabled on the port Ethernet180
20:51:03 ptf.create_macsec_info L0043 INFO | MACsec isn't enabled on the port Ethernet8
20:51:04 ptf.create_macsec_info L0043 INFO | MACsec isn't enabled on the port Ethernet216
..............
20:52:20 ptf.create_macsec_info L0043 INFO | MACsec isn't enabled on the port Ethernet20
20:52:21 ptf.create_macsec_info L0043 INFO | MACsec isn't enabled on the port Ethernet24
20:52:22 ptf.create_macsec_info L0043 INFO | MACsec isn't enabled on the port Ethernet252
logic:
When the test calls to the ptf test, meaning that there will be no more calls to the switch at this time (no additional CLI commands and etc)
For example the test ecmp/inner_hashing/test_wr_inner_hashing.py :
The test doing the warm-reboot and running the ptf test in parallel. Not expected that ptf test will start only in 1+ minutes and before will do some actions on the switch, when the switch doing the warm-reboot. As a result we getting the unreachable switch:
# FIXME - if multiple hosts were involved, should an exception be raised?
if cb.unreachable:
> raise AnsibleConnectionFailure("Host unreachable", dark=cb.unreachable, contacted=cb.contacted)
E AnsibleConnectionFailure: Host unreachable
The text was updated successfully, but these errors were encountered:
Description
By PR #5490 was introduced the degradation of ptf tests runtime and logic.
Before the start of each ptf test, executed the checker for MACsec for each interface.
This step takes ~1:20 minutes.
For example the test ecmp/test_fgnhg.py:
Previously runtime of the test was ~24 minutes.
Now, the test terminated in our regression with a timeout of 1h. Where more than 30 times called the ptf test. It means that more than 30*1:20 =40 minutes from 1 hour we lost on MAcsec checker.
When the test calls to the ptf test, meaning that there will be no more calls to the switch at this time (no additional CLI commands and etc)
For example the test ecmp/inner_hashing/test_wr_inner_hashing.py :
The test doing the warm-reboot and running the ptf test in parallel. Not expected that ptf test will start only in 1+ minutes and before will do some actions on the switch, when the switch doing the warm-reboot. As a result we getting the unreachable switch:
The text was updated successfully, but these errors were encountered: