-
Notifications
You must be signed in to change notification settings - Fork 267
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
Enable RAID support for FDP IO path #319
base: main
Are you sure you want to change the base?
Conversation
cachelib/navy/common/Device.cpp
Outdated
// As of now, only one FDP enabled Device is supported | ||
static constexpr uint16_t kDefaultFdpIdx = 0u; | ||
// Map of file descriptors to FdpNvme device objects | ||
const std::unordered_map<int, std::shared_ptr<FdpNvme>> fdpNvmeDevs_; |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should this be a reference?
I thought all the AsyncIoContext would share the same map (vector before this change)? Or we actually want to make a copy here?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Should this be a reference? I thought all the AsyncIoContext would share the same map (vector before this change)? Or we actually want to make a copy here?
yes, you are right, we can use reference here.
cachelib/navy/common/Device.cpp
Outdated
return fdpNvmeVec_[kDefaultFdpIdx]->allocateFdpHandle(); | ||
if (fdpNvmeDevs_.size() > 0) { | ||
auto fdpHandle = -1; | ||
//Ensuring that same FDP placement handle is allocated for all FdpNvme |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Asking to learn: I thought it would be fine if one RAID write gets written to different handles in different device. But because we don't want to keep track of a vector of different handles, we just make sure we have the same handle for all devices?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Asking to learn: I thought it would be fine if one RAID write gets written to different handles in different device. But because we don't want to keep track of a vector of different handles, we just make sure we have the same handle for all devices?
Yes, we wanted to avoid the code complexity and use a simple scheme of same handle for all the devices for a given placement handle. Having ability to write to different handles for different RAID devices could be nice but need to design carefully, and can be added later.
Summary: This enables RAID0 in fdp io path by spliting io across all devices. Signed-off-by: Vikash Kumar <[email protected]>
1a07b4e
to
40f1b33
Compare
This PR enables RAID in the FDP io path. RAID support was disabled in the original FDP support PR, since it involved a few more changes in the IO Path code. It also involves handling a few scenarios like one of the devices in the RAID does not support FDP etc. This PR contain the code changes for the same.