Backport net.box conn leak and sync test #521
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This commit introduces test, which checks, that vshard properly works
with synchronous user spaces. It also checks, that vshard properly finds
masters, if raft failover is used.
With elections and auto master enabled vshard works as expected,
no additional actions are required.
However, in order to use synchro replication in vshard without
election enabled user is supposed to manually call box.ctl.promote
on the expected master:
```lua
box.cfg{read_only = false}
box.ctl.promote()
```
Note, that `on_master_enable` cannot be used for `box.ctl.promote`,
when the node, which was a master, fails. It was decided to go with
`box.info.ro` flag as the marker of the storage state switch:
replica -> master and vice versa. All `on_master_enable/disable`
callbacks are executed only after `box.info.ro` is changed,
so it's impossible to call `box.ctl.promote` in them, since
instance doesn't go into rw state after box.cfg reconfiguration
due to `synchro` reason.
Closes tarantool#413
NO_DOC=<on_master_enable/disable are not documented>
This commit adds simple load generator for vshard. Currently it generates only different echo requests. In the future support for adding/removing customers and accounts may be introduced, but it's not required for now. NO_DOC=example NO_TEST=example
There was a typo in `netbox_is_conn_dead` function, it marked the connection as dead, when the fiber was in `suspended` status. The connection is down, when its fiber is only in `dead` status, in other cases `reconnect_after` works. Closes tarantool#518 NO_DOC=bugfix
The net.box connection could not be garbage collected until vconnect.future async call is garbage collected, since the connection is referenced as a part of its async request. The leak is reproducable only, when name_as_key identification is used, since in other cases vconnect is not set. Let's remove the future object from the net.box connection on disconnect, so that it can be garbage collected after detach. Part of tarantool#517 NO_DOC=bugfix
When replica is removed on reconfig, then if it has vconnect in progress, it cannot be garbage collected due to vconnect.future async call, which references connection. Let's detach connections of the removed replicas on reconfig, on detach the future object is dropped. Closes tarantool#517 NO_DOC=bugfix
sergepetrenko
requested review from
sergepetrenko
and removed request for
Gerold103
sergepetrenko
approved these changes
Mar 11, 2025
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
No description provided.