Reparenting endtoend testcases in Go migrated from Python [+ github actions] #5604
Conversation
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
This reverts commit d03d3a1. Signed-off-by: Ajeet jain <ajeet@planetscale.com>
ajeetj
force-pushed
the
tal_reparent_new
branch
from
4a5ae73
to
1602423
Compare
| tmClient *tmc.Client | ||
| keyspaceName = "ks" | ||
| shardName = "0" | ||
| shard1Name = "0000000000000000-ffffffffffffffff" |
There was a problem hiding this comment.
Is there a reason to use this instead of just 0?
There was a problem hiding this comment.
No, copied from Python ones. Will test with 0-f
There was a problem hiding this comment.
Updated it to -80. Let me know if that works for the test
There was a problem hiding this comment.
I looked at the original python test case. I think this does need to be the same as it was in python.
Also, in the python test, we basically run the same scenario for shard0 and shard1, the only difference is how the shard is represented: 0 vs 0000...-ffff.... We should preserve that while porting rather than changing the test to only have 1 master + 1 replica when the shard is represented as 0000...-ffff...
There was a problem hiding this comment.
I have reverted the range to 000...-ffff....
As wrt to replica counts I have only used master-replica as we just need to verify replication in a replica in range-based test.
As opposed to other Tests like TestChangeTypeSemiSync and TestReparentFromOutside need 4 replicas setup which is already present in Shard0.
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
ajeetj
changed the title
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
ajeetj
force-pushed
the
tal_reparent_new
branch
from
715e1b6
to
1ec88e1
Compare
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
|
@deepthi This is ready for another review. |
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
ajeetj
changed the title
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
|
LGTM I have changed cluster_endtoend test to not be required. |
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
Signed-off-by: Ajeet jain <ajeet@planetscale.com>
WaitForTabletTypesnew function in vttablet_process to check if tablet status is any of the give statuses. [This was done as we are stopping and starting the same tablet instances, so their status were unpredictable. Sometimes few of them comes in SERVING mode while others comes in NOT_SERVING mode]@deepthi This is ready for review and can be merged after all checks are passed.
Signed-off-by: Ajeet jain ajeet@planetscale.com