feat: Implement MSETEX command #6334

romange · 2026-01-06T15:42:54Z

Adds MSETEX command that atomically sets multiple string keys with a shared expiration time, extending MSET with expiration options similar to SET.

Syntax:
MSETEX numkeys key value [key value ...] [EX seconds|PX milliseconds|EXAT timestamp|PXAT timestamp|KEEPTTL]

Returns:
1 if all keys were set
0 if operation failed

Note: NX/XX options are intentionally not supported as ensuring atomicity for these conditional operations is much harder in Dragonfly's multi-threaded architecture. These options will return "unsupported option" error.

Implementation:

OpMSetEx: Handles per-shard operations with expiration options
CmdMSetEx: Parses command arguments and coordinates transaction
DetermineKeys: Extended to handle VARIADIC_KEYS with INTERLEAVED_KEYS
Added bounds validation for interleaved key-value arguments

Testing:

Comprehensive unit tests covering all expiration options (EX, PX, EXAT, PXAT, KEEPTTL)
Tests for NX/XX rejection with unsupported option error
Error handling tests for invalid arguments

Fixes #6329

augmentcode · 2026-01-06T15:48:34Z

🤖 Augment PR Summary

Summary: Introduces the Redis-compatible MSETEX command to atomically set multiple string keys while applying a shared expiration policy.

Changes:

Added CmdMSetEx/OpMSetEx to parse, shard, and execute multi-key sets with TTL options.
Supported expiration modifiers similar to SET: EX, PX, EXAT, PXAT, and KEEPTTL.
Explicitly rejects NX/XX for now due to atomicity constraints in the multi-threaded/sharded model.
Added ParseExpireTime helper and reused it in (P)SETEX parsing.
Extended DetermineKeys to correctly handle VARIADIC_KEYS + INTERLEAVED_KEYS key/value layouts, with bounds validation.
Adjusted transaction argument validation to use KeyIndex::NumArgs() for interleaved commands.
Added unit tests covering success paths, TTL options, invalid arguments, and unsupported options.

Technical Notes: Execution is coordinated via a single-hop transaction and uses explicit journaling to replicate the resulting per-key updates (including TTL via PXAT/KEEPTTL).

_{🤖 Was this summary useful? React with 👍 or 👎}

augmentcode

Review completed. 2 suggestions posted.

Comment augment review to trigger a new review at any time.

src/server/string_family.cc

Copilot

Pull request overview

This PR implements the MSETEX command, which atomically sets multiple string keys with a shared expiration time, extending MSET functionality with expiration options similar to SET. The command syntax is MSETEX numkeys key value [key value ...] [EX seconds|PX milliseconds|EXAT timestamp|PXAT timestamp|KEEPTTL], returning 1 on success or 0 on failure. A notable deviation from the original issue is that NX/XX options are intentionally not supported due to atomicity constraints in Dragonfly's multi-threaded architecture.

Key changes:

Extended transaction key determination logic to properly handle VARIADIC_KEYS with INTERLEAVED_KEYS patterns
Added ParseExpireTime helper function for consistent expiration time validation across commands
Implemented OpMSetEx and CmdMSetEx for the core MSETEX functionality with explicit journaling

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 4 comments.

File	Description
src/server/transaction.cc	Extended DetermineKeys to support VARIADIC_KEYS with INTERLEAVED_KEYS by adjusting end calculation and adding bounds validation; updated DCHECK to use NumArgs() for correct assertion with interleaved keys
src/server/string_family.cc	Added ParseExpireTime helper, OpMSetEx operation handler with explicit journaling, CmdMSetEx command implementation with expiration option parsing, and registered MSETEX command with VARIADIC_KEYS and INTERLEAVED_KEYS flags
src/server/string_family_test.cc	Added comprehensive MSetEx test covering basic usage, expiration options (EX, PX, KEEPTTL), error cases, and verification that NX/XX options return unsupported option errors

src/server/string_family.cc

Copilot · 2026-01-06T15:51:07Z

src/server/string_family_test.cc

+TEST_F(StringFamilyTest, MSetEx) {
+  // Basic usage - sets two keys with expiration
+  EXPECT_EQ(1, CheckedInt({"msetex", "2", "key1", "val1", "key2", "val2", "EX", "100"}));
+  EXPECT_EQ(Run({"get", "key1"}), "val1");
+  EXPECT_EQ(Run({"get", "key2"}), "val2");
+  EXPECT_THAT(Run({"ttl", "key1"}), IntArg(100));
+  EXPECT_THAT(Run({"ttl", "key2"}), IntArg(100));
+
+  // Basic usage without expiration
+  EXPECT_EQ(1, CheckedInt({"msetex", "1", "key3", "val3"}));
+  EXPECT_EQ(Run({"get", "key3"}), "val3");
+  EXPECT_THAT(Run({"ttl", "key3"}), IntArg(-1));
+
+  // PX option (milliseconds)
+  EXPECT_EQ(1, CheckedInt({"msetex", "1", "key4", "val4", "PX", "5000"}));
+  EXPECT_EQ(Run({"get", "key4"}), "val4");
+  auto ttl = Run({"pttl", "key4"});
+  EXPECT_GT(get<int64_t>(ttl.u), 4000);
+
+  // KEEPTTL option
+  Run({"set", "ttl_key", "original", "EX", "200"});
+  EXPECT_EQ(1, CheckedInt({"msetex", "1", "ttl_key", "new_value", "KEEPTTL"}));
+  EXPECT_EQ(Run({"get", "ttl_key"}), "new_value");
+  auto ttl_val = Run({"ttl", "ttl_key"});
+  EXPECT_GT(get<int64_t>(ttl_val.u), 0);
+  EXPECT_LE(get<int64_t>(ttl_val.u), 200);
+
+  // Error cases
+  EXPECT_THAT(Run({"msetex", "abc", "k", "v"}), ErrArg("not an integer"));  // invalid numkeys
+  EXPECT_THAT(Run({"msetex", "0", "k", "v"}), ErrArg("not an integer"));    // numkeys <= 0
+  EXPECT_THAT(Run({"msetex", "2", "k", "v"}), ErrArg("syntax"));  // not enough key-value pairs
+  EXPECT_THAT(Run({"msetex", "1", "k", "v", "EX", "10", "PX", "1000"}),
+              ErrArg("syntax"));  // multiple expiration options
+  EXPECT_THAT(Run({"msetex", "1", "k", "v", "EX", "0"}),
+              ErrArg("invalid expire time"));  // zero expiration
+  EXPECT_THAT(Run({"msetex", "1", "k", "v", "EX", "-1"}),
+              ErrArg("invalid expire time"));  // negative expiration
+
+  // NX/XX options are unsupported due to atomicity constraints in multi-threaded execution
+  EXPECT_THAT(Run({"msetex", "1", "k", "v", "NX"}), ErrArg("unsupported option"));
+  EXPECT_THAT(Run({"msetex", "1", "k", "v", "XX"}), ErrArg("unsupported option"));
+  EXPECT_THAT(Run({"msetex", "2", "k1", "v1", "k2", "v2", "NX", "EX", "100"}),
+              ErrArg("unsupported option"));
+}


Test coverage is missing for EXAT and PXAT expiration options. While EX, PX, and KEEPTTL are tested, absolute timestamp expiration options (EXAT and PXAT) should also be tested to ensure they work correctly with MSETEX. Consider adding tests similar to those that exist for the SET command with EXAT/PXAT options.

Copilot · 2026-01-06T15:51:07Z

src/server/string_family.cc

+    auto expire_ms = ParseExpireTime(int_val, is_ms, is_absolute);
+    if (!expire_ms) {
+      return cmnd_cntx->SendError(InvalidExpireTime("msetex"));
+    }
+
+    sparams.flags |= SetCmd::SET_EXPIRE_AFTER_MS;
+    sparams.expire_after_ms = *expire_ms;
+  }


Missing handling for expiration times in the past. When using EXAT or PXAT with a timestamp that has already passed, the SET command (lines 1062-1076) has special logic to delete existing keys and return success. However, MSETEX doesn't handle this case - it will successfully set keys with negative relative expiration times. The keys should either be deleted immediately (like SET does) or the operation should fail with an appropriate error message.

Copilot · 2026-01-06T17:33:43Z

@romange I've opened a new pull request, #6335, to work on those changes. Once the pull request is ready, I'll request review from you.

Adds MSETEX command that atomically sets multiple string keys with a shared expiration time, extending MSET with expiration options similar to SET. Syntax: MSETEX numkeys key value [key value ...] [EX seconds|PX milliseconds|EXAT timestamp|PXAT timestamp|KEEPTTL] Returns: 1 if all keys were set 0 if operation failed Note: NX/XX options are intentionally not supported as ensuring atomicity for these conditional operations is much harder in Dragonfly's multi-threaded architecture. These options will return "unsupported option" error. Implementation: - OpMSetEx: Handles per-shard operations with expiration options - CmdMSetEx: Parses command arguments and coordinates transaction - DetermineKeys: Extended to handle VARIADIC_KEYS with INTERLEAVED_KEYS - Added bounds validation for interleaved key-value arguments Testing: - Comprehensive unit tests covering all expiration options (EX, PX, EXAT, PXAT, KEEPTTL) - Tests for NX/XX rejection with unsupported option error - Error handling tests for invalid arguments Fixes #6329 Signed-off-by: Roman Gershman <[email protected]>

Copilot

Pull request overview

Copilot reviewed 4 out of 4 changed files in this pull request and generated 2 comments.

Copilot · 2026-01-07T06:41:12Z

src/server/string_family.cc

+// The returned value is suitable for SetCmd::SetParams::expire_after_ms.
+optional<uint64_t> ParseExpireTime(int64_t int_val, bool is_ms, bool is_absolute) {
+  DbSlice::ExpireParams expiry{
+      .value = int_val,
+      .unit = is_ms ? TimeUnit::MSEC : TimeUnit::SEC,
+      .absolute = is_absolute,
+  };
+
+  int64_t now_ms = GetCurrentTimeMs();
+  auto [rel_ms, abs_ms] = expiry.Calculate(now_ms, false);
+  // Reject if absolute time is negative or if expiration is already in the past
+  if (abs_ms < 0 || rel_ms < 0) {
+    return nullopt;
+  }
+


The ParseExpireTime function has a bug when handling absolute timestamps (EXAT/PXAT) that are in the past. When an absolute timestamp is provided that is positive but less than the current time, the calculated relative time (rel_msec) will be negative. This negative int64_t value is then implicitly converted to uint64_t when returned, causing an integer overflow that results in a very large positive expiration time instead of rejecting the invalid timestamp or treating it as expired. The function should check if the calculated relative time is negative or zero when absolute timestamps are used and return nullopt in such cases.

Copilot · 2026-01-07T06:41:13Z

src/server/string_family_test.cc

+TEST_F(StringFamilyTest, MSetEx) {
+  // Basic usage - sets two keys with expiration
+  EXPECT_EQ(1, CheckedInt({"msetex", "2", "key1", "val1", "key2", "val2", "EX", "100"}));
+  EXPECT_EQ(Run({"get", "key1"}), "val1");
+  EXPECT_EQ(Run({"get", "key2"}), "val2");
+  EXPECT_THAT(Run({"ttl", "key1"}), IntArg(100));
+  EXPECT_THAT(Run({"ttl", "key2"}), IntArg(100));
+
+  // Basic usage without expiration
+  EXPECT_EQ(1, CheckedInt({"msetex", "1", "key3", "val3"}));
+  EXPECT_EQ(Run({"get", "key3"}), "val3");
+  EXPECT_THAT(Run({"ttl", "key3"}), IntArg(-1));
+
+  // PX option (milliseconds)
+  EXPECT_EQ(1, CheckedInt({"msetex", "1", "key4", "val4", "PX", "5000"}));
+  EXPECT_EQ(Run({"get", "key4"}), "val4");
+  auto ttl = Run({"pttl", "key4"});
+  EXPECT_GT(get<int64_t>(ttl.u), 4000);
+
+  // EXAT option (absolute timestamp in seconds)
+  int64_t future_ts = time(nullptr) + 100;
+  EXPECT_EQ(1, CheckedInt({"msetex", "1", "exat_key", "val", "EXAT", to_string(future_ts)}));
+  EXPECT_EQ(Run({"get", "exat_key"}), "val");
+  auto exat_ttl = Run({"ttl", "exat_key"});
+  EXPECT_GT(get<int64_t>(exat_ttl.u), 90);
+  EXPECT_LE(get<int64_t>(exat_ttl.u), 100);
+
+  // PXAT option (absolute timestamp in milliseconds)
+  int64_t future_ts_ms = (time(nullptr) + 100) * 1000;
+  EXPECT_EQ(1, CheckedInt({"msetex", "1", "pxat_key", "val", "PXAT", to_string(future_ts_ms)}));
+  EXPECT_EQ(Run({"get", "pxat_key"}), "val");
+  auto pxat_ttl = Run({"pttl", "pxat_key"});
+  EXPECT_GT(get<int64_t>(pxat_ttl.u), 90000);
+  EXPECT_LE(get<int64_t>(pxat_ttl.u), 100000);
+
+  // KEEPTTL option
+  Run({"set", "ttl_key", "original", "EX", "200"});
+  EXPECT_EQ(1, CheckedInt({"msetex", "1", "ttl_key", "new_value", "KEEPTTL"}));
+  EXPECT_EQ(Run({"get", "ttl_key"}), "new_value");
+  auto ttl_val = Run({"ttl", "ttl_key"});
+  EXPECT_GT(get<int64_t>(ttl_val.u), 0);
+  EXPECT_LE(get<int64_t>(ttl_val.u), 200);
+
+  // Error cases


The test suite is missing coverage for EXAT and PXAT options which are mentioned in the command syntax and implementation. Tests should be added to verify that absolute timestamps work correctly, including edge cases like timestamps in the future, timestamps around the current time, and invalid timestamps (e.g., timestamps in the past).

Copilot AI review requested due to automatic review settings January 6, 2026 15:42

Copilot started reviewing on behalf of romange January 6, 2026 15:43 View session

romange force-pushed the AddMSetEx branch from 7210ee1 to 29af7b5 Compare January 6, 2026 15:44

augmentcode bot reviewed Jan 6, 2026

View reviewed changes

src/server/string_family.cc Show resolved Hide resolved

src/server/string_family.cc Show resolved Hide resolved

Copilot AI reviewed Jan 6, 2026

View reviewed changes

romange force-pushed the AddMSetEx branch from 29af7b5 to b250dc0 Compare January 6, 2026 17:32

Copilot AI mentioned this pull request Jan 6, 2026

Fix partial failure handling in OpMSetEx #6335

Closed

romange force-pushed the AddMSetEx branch from b250dc0 to 680f50a Compare January 7, 2026 06:34

Copilot AI review requested due to automatic review settings January 7, 2026 06:34

Copilot started reviewing on behalf of romange January 7, 2026 06:34 View session

romange force-pushed the AddMSetEx branch from 680f50a to f20a21d Compare January 7, 2026 06:39

Copilot AI reviewed Jan 7, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: Implement MSETEX command #6334

feat: Implement MSETEX command #6334

romange commented Jan 6, 2026

Uh oh!

augmentcode bot commented Jan 6, 2026

Uh oh!

augmentcode bot left a comment

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Uh oh!

Uh oh!

Uh oh!

Copilot AI Jan 6, 2026

Uh oh!

Copilot AI Jan 6, 2026

Uh oh!

Copilot AI commented Jan 6, 2026

Uh oh!

Copilot AI left a comment

Uh oh!

Copilot AI Jan 7, 2026

Uh oh!

Copilot AI Jan 7, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

feat: Implement MSETEX command #6334

Are you sure you want to change the base?

feat: Implement MSETEX command #6334

Conversation

romange commented Jan 6, 2026

Uh oh!

augmentcode bot commented Jan 6, 2026

Uh oh!

augmentcode bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

Uh oh!

Uh oh!

Copilot AI Jan 6, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 6, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI commented Jan 6, 2026

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Uh oh!

Copilot AI Jan 7, 2026

Choose a reason for hiding this comment

Uh oh!

Copilot AI Jan 7, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants