DOCS-14695-Upgrading language strength with regards to hidden delayed… #5787
Conversation
| and querying delayed data without a direct connection to that member. | ||
| Making delayed replica set members non-voting means they will not | ||
| count towards acknowledging write operations with read concern | ||
| :readconcern:`"majority"`. |
There was a problem hiding this comment.
Should this be "write concern majority" ? Instead of read concern majority?
There was a problem hiding this comment.
Updated to write concern.
| If you do not hide delayed members and one or more nodes | ||
| becomes unavailable, the replica set has to wait for the delayed | ||
| member and the commit point lags. A lagged commit point can lead to | ||
| performance issues. |
There was a problem hiding this comment.
There is some shorthand and tone in this section and following example that doesn't really match the tone of the docs. We don't really use "config" as shorthand for "configuration", or "mins" as short for "minutes".
Consider something like:
For example, consider a Primary-Secondary-Delayed replica set
configuration where the delayed secondary is voting with a 10
minute delay.
With one non-delayed secondary unavailable, the degraded configuration
of Primary-Delayed must wait at least 10 minutes to acknowledge a write
operation with :writeconcern:`"majority"`.
The majority commit point will take longer to advance, leading to cache
pressure similar performance issues with...
| - Primary-Secondary-Delayed replica set config where delayed | ||
| secondary is voting with 10 mins delay. |
There was a problem hiding this comment.
This bullet is a sentence fragment which reads a little awkwardly. See suggestion in the comment above.
| - The majority commit point will take longer to advance leading to | ||
| cache pressure similar to performance issues with a PSA replica | ||
| set. |
There was a problem hiding this comment.
Can we end this thought with something like:
For more information on the majority commit point, see Causal Consistency and Read and Write Concerns
I think "majority commit point" is a little confusing and providing that link for context may be useful.
I'm also not sure I understand the bit about a PSA replica set. We haven't mentioned them at all in this page and I feel like I'm missing some context. Is there anywhere we can link to in the docs that explains more about these performance issues? It kind of seems to come out of nowhere. We should also expand the acronym since users might not know what we mean by a PSA replica set.
There was a problem hiding this comment.
Updated / added a link to the replica set maintenance tutorial, which goes into more context of resolving the performance issues detailed with the commit point.
|
@ianf-mongodb , just pinging in case these changes fell off your radar. |
ianf-mongodb
force-pushed
the
DOCS-14695-hidden-replica-read-operations-verbiage
branch
3 times, most recently
from
240fcca to
e2a50b3
Compare
| @@ -39,12 +39,14 @@ Delayed members: | |||
| members. Set the priority to 0 to prevent a delayed member from | |||
| becoming primary. | |||
|
|
|||
| - **Should be** :ref:`hidden <replica-set-hidden-members>` | |||
| - **Must be** :ref:`hidden <replica-set-hidden-members>` | |||
| members. Always prevent applications from seeing and querying | |||
| delayed members. | |||
|
|
|||
| - *Do* vote in :term:`elections <election>` for primary, if :rsconf:`members[n].votes` is set to 1. | |||
There was a problem hiding this comment.
I'm a little confused because here we are saying that the delayed members vote in elections for the primary, but then immediately below we say that users should ensure that they are non-voting. It seems slightly contradictory.
Do you think we should we should add a sentence to this bullet that says something like:
Ensuring that delayed members are non-voting by
setting :rsconf:`members[n].votes` to 0 can help improve performance.
There was a problem hiding this comment.
Im reading this as a warning to the user, in other words the first two bullets, which start with Must be are the recommendation. The Do is a warning if you do not address the rsconfigmembers[n].votes setting. I think the extra blurb does add value.
|
|
||
| For more information on the majority commit point, see | ||
| :doc:`Causal Consistency and Read and Write Concerns | ||
| </core/causal-consistency-read-write-concerns>` for additional details |
There was a problem hiding this comment.
This should be two separate sentences?
There was a problem hiding this comment.
Separated in latest.
ianf-mongodb
force-pushed
the
DOCS-14695-hidden-replica-read-operations-verbiage
branch
from
6ca41ec to
14a9c01
Compare
| With one non-delayed secondary unavailable, the degraded configuration | ||
| of Primary-Delayed must wait at least 10 minutes to acknowledge a write | ||
| operation with :writeconcern:`"majority"`. | ||
|
|
||
| The majority commit point will take longer to advance, leading to cache | ||
| pressure similar performance issues with a | ||
| :ref:`Primary with a Secondary and an Arbiter<rs-architecture-psa>` | ||
| (PSA) replica set. |
There was a problem hiding this comment.
Can we combine these into a single paragraph? This block is very big because we have so many paragraphs so I'm looking for ways to combine some of them.
| For more information on the majority commit point, see | ||
| :doc:`Causal Consistency and Read and Write Concerns | ||
| </core/causal-consistency-read-write-concerns>`. | ||
|
|
||
| For additional details on resolving performance issues see the | ||
| :ref:`replica set maintenance tutorial<performance-issues-psa>`. |
There was a problem hiding this comment.
Similarly, we could combine these into a single paragraph as well.
ianf-mongodb
force-pushed
the
DOCS-14695-hidden-replica-read-operations-verbiage
branch
from
14a9c01 to
b56d29a
Compare
* DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn --------- Co-authored-by: jason-price-mongodb <[email protected]>
* DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn --------- Co-authored-by: jason-price-mongodb <[email protected]>
* DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn --------- Co-authored-by: jason-price-mongodb <[email protected]>
* DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn --------- Co-authored-by: jason-price-mongodb <[email protected]>
* DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn * DOCSP-35359-6.0.13-rn --------- Co-authored-by: jason-price-mongodb <[email protected]>
No description provided.