-
Notifications
You must be signed in to change notification settings - Fork 4.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
max_open_connections default value is too low for the oracle database plugin #6790
Comments
Happy to bump up the default from 2 (after all ideally this just works), but you can already override the default per mount -- https://www.vaultproject.io/api/secret/databases/oracle.html#max_open_connections |
I understand you can override the value on the mount configuration. That works fine, but if you leave it on the default value for the oracle plugin, every mount you don't set this configuration on can get a deadlock waiting for an available connection. I was thinking that bumping up the default for all database plugins might not be acceptable as a solution, so allowing each plugin to set the default would be an option. If a higher default value is acceptable for all database plugins, then that's definitely the simplest solution. |
What would a good default value be if we were to bump it up? |
@mjarmy I think 4 would be good for now. If we go too high I'd be concerned that people might not be prepared for a potential tripling of connections. 4 across all database types as the new default seems sane, and we can document in CHANGES. |
I've posted a draft PR that addresses this: #7093 |
This is done, since #7093 has been merged |
Is your feature request related to a problem? Please describe.
The
max_open_connections
setting for database plugins has a default value of2
, which is too low for the oracle database plugin. See hashicorp/vault-plugin-database-oracle#16Describe alternatives you've considered
I see two options:
3
. However, I've seen issues related to too many file descriptors ("Too many open files" #1257 for instance) that reference this setting. So this solution could increase instances of such problems for users who aren't explicitly setting this value already.Describe the solution you'd like
I don't have a preference. I expect 1 is easier, but 2 will have a less wide-spread impact and allow opt-in rather than opt-out.
Feedback from those who have a better sense of the repercussions of one over the other would be nice.
The text was updated successfully, but these errors were encountered: