DAOS-18381 ddb: properly transfer ddb cmdline options #17312
Conversation
|
Ticket title is 'ddb rm_pool failed to initialize sysdb under md-on-ssd mode' |
|
Test stage NLT on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17312/1/testReport/ |
|
Test stage Functional on EL 8.8 completed with status UNSTABLE. https://jenkins-3.daos.hpc.amslabs.hpecorp.net/job/daos-stack/job/daos//view/change-requests/job/PR-17312/1/testReport/ |
The db_path maybe not contained inside the vos_path parameter, especially under md-on-ssd mode. Related go interfaces need to properly transfer cmdline options (db_path, write_mode, and so on) from control plane to the lower ddb utils. The patch also adds more check for the vos pool open and close status before real VOS operation. Features: control recovery Signed-off-by: Fan Yong <fan.yong@hpe.com>
Nasf-Fan
force-pushed
the
Nasf-Fan/DAOS-18381_test
branch
from
388b194 to
b872331
Compare
| param := []string{} | ||
| skip := true | ||
|
|
||
| for _, v := range args { | ||
| if v == string(opts.Args.VosPath) && skip { | ||
| skip = false | ||
| } else if v != string(opts.Args.RunCmd) { | ||
| param = append(param, v) | ||
| } | ||
| } |
There was a problem hiding this comment.
Why does VosPath need to be removed from the list of params? Would be good to have a brief comment here documenting the reason.
There was a problem hiding this comment.
If not filter out VosPath, then for non-opened sub-command, such as rm_pool will report failure, for example:
ERROR: ddb: invalid usage of command 'rm_pool' (unconsumed input '/mnt/nasf_1/b9eac3f1-1c3d-44e9-bf1d-285bea2402a8/vos-0'), try 'help'
I prefer to update the comment in another PR to avoid triggering CI test for this one since current master CI test is not stable. How do you think? @kjacque
There was a problem hiding this comment.
OK, I'm fine with that. I see that this is related to DAOS-18382. This gives me a better understanding of the issue, thanks.
There was a problem hiding this comment.
Thanks @kjacque , we will land my patch firstly, then Makito can continue his CR test against the latest master.
And then, apply yours when related control plane patch is ready.
The db_path maybe not contained inside the vos_path parameter, especially under md-on-ssd mode. Related go interfaces need to properly transfer cmdline options (db_path, write_mode, and so on) from control plane to the lower ddb utils.
The patch also adds more check for the vos pool open and close status before real VOS operation.
Features: control recovery
Steps for the author:
After all prior steps are complete: