aboutsummaryrefslogtreecommitdiff
path: root/src/zenstore
Commit message (Collapse)AuthorAgeFilesLines
* reduce memory churn (#248)Stefan Boberg2024-12-021-5/+5
| | | | | * eliminated allocation in SetCurrentThreadName * reduced memory allocator activity in cache RPC response building * reduced allocations in compact binary building
* validate that root path exists for workspace before adding it (#250)Dan Engelbrecht2024-12-021-0/+5
|
* added support for dynamic LLM tags (#245)Stefan Boberg2024-12-028-0/+191
| | | | | * added FLLMTag which can be used to register memory tags outside of core * changed `UE_MEMSCOPE` -> `ZEN_MEMSCOPE` for consistency * instrumented some subsystems with dynamic tags
* make sure we don't throw exception from worker thread (#247)Dan Engelbrecht2024-11-281-3/+17
| | | | | | * Make sure we don't throw exception from worker thread * secure async project flush * secure workspaces * spelling
* use plain sorted array instead of map of vectors (#237)Dan Engelbrecht2024-11-272-22/+33
| | | | | * use plain sorted array instead of map of vectors * reserve vectors up front = 5% perf increase * don't do batch read of chunks if we have a single chunk -> 1% perf gain
* add missing shard lock in filecas compact (#239)Dan Engelbrecht2024-11-271-1/+6
| | | | * add missing shardlock during compact in filecas * add warning log when filecas fails to open a file it expects to be present
* caller controls threshold for bulk-loading chunks in IterateChunks (#222)Dan Engelbrecht2024-11-259-18/+31
| | | | | | * Allow caller to control threshold for bulk-loading chunks in IterateChunks * use smaller batch chunk reading for /fileinfos and /chunkinfos as we do not intend to read the payload * use smaller batch read buffer when just querying for size of attachments
* stronger validation of payload existance (#229)Dan Engelbrecht2024-11-253-57/+87
| | | | | | - Don't add RawSize and Size in ProjectStore::GetProjectFiles response if we can't get the payload - Use validation of payload size/existance in all chunk fetch operations in file cas - In project store oplog validate, make sure we can reach all the payloads - Add threading to oplog validate request
* add missing projectstore expire time in gc log (#227)Dan Engelbrecht2024-11-251-0/+1
|
* split zencore/memory.h -> memoryview.h, memcmp.h (#228)Stefan Boberg2024-11-252-2/+2
| | | | | | | minor clean-up `zencore/memory.h` used to contain a variety of things including `Malloc` support along with `MemoryView` etc since the memory allocator stuff moved into `zencore/memory/memory.h` there was basically only `MemoryView` and `MemCmp` in there which seemed better to split out into separate headers to avoid overloading `memory.h`
* Insights-compatible memory tracking (#214)Stefan Boberg2024-11-251-1/+1
| | | | | | | | | | | | | This change introduces support for tracing of memory allocation activity. The code is ported from UE5, and Unreal Insights can be used to analyze the output. This is currently only fully supported on Windows, but will be extended to Mac/Linux in the near future. To activate full memory tracking, pass `--trace=memory` on the commandline alongside `--tracehost=<ip>` or `-tracefile=<path>`. For more control over how much detail is traced you can instead pass some combination of `callstack`, `memtag`, `memalloc` instead. In practice, `--trace=memory` is an alias for `--trace=callstack,memtag,memalloc`). For convenience we also support `--trace=memory_light` which omits call stacks. This change also introduces multiple memory allocators, which may be selected via command-line option `--malloc=<allocator>`: * `mimalloc` - mimalloc (default, same as before) * `rpmalloc` - rpmalloc is another high performance allocator for multithreaded applications which may be a better option than mimalloc (to be evaluated). Due to toolchain limitations this is currently only supported on Windows. * `stomp` - an allocator intended to be used during development/debugging to help track down memory issues such as use-after-free or out-of-bounds access. Currently only supported on Windows. * `ansi` - fallback to default system allocator
* fix inconsistencies in filecas due to failing to remove payload file during ↵Dan Engelbrecht2024-11-223-427/+436
| | | | | | | | GC (#224) make sure we rewrite filecas entries if chunk size changes (due to compression changes) hardening of move/write files in filecas if we encounter a filecas entry with mismatching size (due to pre-existing bug) we validate the file and update the index if we find a bad filecas file on disk we now attempt to remove it
* oplog prep gc fix (#216)Dan Engelbrecht2024-11-156-239/+525
| | | | | | - Added option gc-validation to zenserver that does a check for missing references in all oplog post full GC. Enabled by default. - Feature: Added option gc-validation to zen gc command to control reference validation. Enabled by default. - Added more details in post GC log. - Fixed race condition in oplog writes which could cause used attachments to be incorrectly removed by GC
* memory leak fix (GetBatchHandle) (#215)Stefan Boberg2024-11-081-0/+4
| | | This fixes a memory leak which would cause stale handles to accumulate until process shutdown. Each cache get operation would therefore leak some memory.
* workspace share security (#192)Dan Engelbrecht2024-10-232-383/+890
| | | | | | | - Improvement: Reworked workspace shares to be more secure. Workspaces and workspace shares can only be created using the `zen workspace` command, the http endpoint is disabled unless zenserver is started with the `--workspaces-allow-changes` option enabled. - Each workspace are now configured via a `zenworkspaceconfig.json` file in the root of each workspace - A workspace can allow shares to be created via the http interface if the workspace is created with the `--allow-share-create-from-http` option enabled - A new http endpoint at `/ws` - issuing a `Get` operation will get you a list of workspaces - A new http endpoint at `/ws/refresh` - issuing a `Get` will make zenserver scan for edits in workspaces and workspace shares
* Use a smaller thread pool during pre-cache phase of GC to reduce memory ↵Dan Engelbrecht2024-10-221-7/+11
| | | | pressure (#205)
* bucket size queries (#203)Dan Engelbrecht2024-10-215-14/+106
| | | - Feature: Added options --bucketsize and --bucketsizes to zen cache-info to get data sizes in cache buckets and attachments
* validate compact binary rpc requests before trying to parse them (#200)Dan Engelbrecht2024-10-161-1/+10
|
* safer path from handle (#195)Dan Engelbrecht2024-10-161-31/+41
| | | * remove PathFromHandle that throws to give better context on failures
* don't read chunks into memory during cache batch fetch unless we may cache ↵Dan Engelbrecht2024-10-092-13/+14
| | | | | them in memory (#188) * Don't read chunks into memory during cache batch fetch unless we may cache them in memory
* remove gc v1 (#121)Dan Engelbrecht2024-10-0314-2819/+99
| | | | | * kill gc v1 * block use of gc v1 from zen command line * warn and flip to gcv2 if --gc-v2=false is specified for zenserver
* gc block size target max size (#180)Dan Engelbrecht2024-10-024-42/+67
| | | | | | * If a block is small (less than half max size) we add it to blocks to compact Sort blocks when iterating over them * do compact of block stores even if no new unused are found * do compact phase even if bucket is empty
* Porject -> ProjectStefan Boberg2024-10-021-1/+1
|
* optimize gc reference sort (#179)Dan Engelbrecht2024-10-012-37/+40
| | | | | - Do a single call to mempcy when fetching attachments from the meta store in GC - Use small lambda when calling std::sort in FilterReferences (enables inlining of the comparision function) - Use a single function for < and == comparision in KeepUnusedReferences
* use alternate IoHash comparision function (#177)v5.5.8-pre5Dan Engelbrecht2024-09-301-4/+23
| | | * Use alternate IoHash comparision function - reduces KeepUnusedReferences execution time by ~20%
* gc command attachment options (#176)Dan Engelbrecht2024-09-304-17/+48
| | | * zen command - add options to control meta data cache when triggering gc
* optimize startup time (#175)Dan Engelbrecht2024-09-304-35/+26
| | | | | | * use tsl::robin_set for BlockIndexSet don't calculate full block location when only block index is needed * don't copy visitor function * reserve space for attachments
* reduce lock time for memcache trim (#171)Dan Engelbrecht2024-09-272-104/+129
| | | | | | | - Improvement: Faster memcache trimming - Reduce calculations while holding bucket lock for memcache trim analysis to reduce contention - When trimming memcache, evict 25% more than required to reduce frequency of trimming - When trimming memcache, don't repack memcache data vector, defer that to regular garbage collection - When trimming memcache, deallocate memcache buffers when not holding exclusive lock in bucket
* Add `gc-attachment-passes` option to zenserver (#167)Dan Engelbrecht2024-09-256-45/+166
| | | | | Added option `gc-attachment-passes` to zenserver Cleaned up GCv2 start and stop logs and added identifier to easily find matching start and end of a GC pass in log file Fixed project store not properly sorting references found during lock phase
* exception safety when writing block (#168)Dan Engelbrecht2024-09-251-10/+8
| | | | * make sure we always clear writing block from m_ActiveWriteBlocks even if we have an exception
* Added namespace qualifier (optional) for z$ rpc requests (#166)Stefan Boberg2024-09-232-1/+17
| | | This change adds support for a namespace-qualified RPC endpoint for z$ at `/z$/<namespace>/$rpc` which may be used to validate RPC requests by URL inspection. The old scheme is still supported.
* gc unused refactor (#165)Dan Engelbrecht2024-09-237-85/+324
| | | | | * optimize IoHash and OId comparisions * refactor filtering of unused references * add attachment filtering to gc
* made fmt formatter format function const (#162)Stefan Boberg2024-09-201-1/+1
| | | this appears to be required as of fmt v11
* unblock PreCache (#164)Dan Engelbrecht2024-09-203-7/+9
| | | Don't lock disk cache buckets from writing when scanning records for attachment references
* gc performance improvements (#160)Dan Engelbrecht2024-09-175-96/+117
| | | | | | | | | | * optimized ValidateCbUInt * optimized iohash comparision * replace unordered set/map with tsl/robin set/map in blockstore * increase max buffer size when writing cache bucket sidecar * only store meta data for files < 4Gb * faster ReadAttachmentsFromMetaData * remove memcpy call in BlockStoreDiskLocation * only write cache bucket state to disk if GC deleted anything
* trace scopes improvementsDan Engelbrecht2024-09-103-9/+11
|
* don't add batch overhead if we are only going to put one cache valueDan Engelbrecht2024-09-101-5/+19
|
* only add mem cached items to list to fill metadata if we don't have metadata ↵Dan Engelbrecht2024-09-101-40/+39
| | | | (#151)
* fix race condition in zenserver during batched fetch (#149)Dan Engelbrecht2024-09-091-1/+92
| | | * fix race condition in zenserver duing batched fetch
* clean cache slog files on startup (#143)Dan Engelbrecht2024-09-042-22/+38
| | | | - Bugfix: If we fail to move a temporary file into place, try to re-open the file so we clean it up - Improvement: Clean up cache bucket log files at startup as we store the matching information in the index snapshot for the bucket
* move gc logs to gc logger (#142)Dan Engelbrecht2024-09-046-2/+38
| | | - Improvement: Move GC logging in callback functions into "gc" context
* oplog index snapshots (#140)Dan Engelbrecht2024-09-034-4/+8
| | | - Feature: Added project store oplog index snapshots for faster opening of oplog - opening oplogs are roughly 10x faster
* meta info store (#75)Dan Engelbrecht2024-08-304-30/+284
| | | | - Feature: Added option `--gc-cache-attachment-store` which caches referenced attachments in cache records on disk for faster GC - default is `false` - Feature: Added option `--gc-projectstore-attachment-store` which caches referenced attachments in project store oplogs on disk for faster GC - default is `false`
* Make sure `noexcept` functions does not leak exceptions (#136)Dan Engelbrecht2024-08-232-168/+208
|
* separate worker pools into burst/background to avoid background jobs ↵Dan Engelbrecht2024-08-223-5/+5
| | | | blocking client requests (#134)
* remove bad assert when payload is memcached but metadata is not set (#130)Dan Engelbrecht2024-08-221-1/+0
|
* close payload file if size mismatch for file cas (#128)Dan Engelbrecht2024-08-201-2/+2
|
* don't throw exception if sidecar file is missing (#126)Dan Engelbrecht2024-08-191-1/+14
| | | | * Don't throw exception if sidecar file is missing. * Log reasons for rejecting sidecar file
* if disk space is low, set the last gc time to avoid spamming retries (#124)Dan Engelbrecht2024-08-191-0/+2
| | | * if disk space is low, set the last gc time to avoid spamming retries
* If we fail to get compression info for a partial request - log and report a ↵Dan Engelbrecht2024-08-191-6/+18
| | | | | miss (#122) * If we fail to get compression info for a partial request - log and report a miss
generated by cgit v1.2.3 (git 2.25.1) at 2026-04-23 12:41:03 +0000