Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
[FAB-5497] Create 1.0.1 fabric release
Single commit which squashes commits from master for 1.0.1 release based on FAB-5468 Change-Id: I4bb6c946d28ba2009c4074c76b4d1451b3d9a1ad Signed-off-by: Gari Singh <[email protected]> [FAB-5280] Fix git clone fabric-samples for Windows This change expands the prerequisites documentation for Windows to instruct users on setting git properly to ensure it handles long filenames (required to successfully clone fabric-samples) and end-of-lines (required for vagrant to work). Change-Id: I8230f1980fedfb43d6e9cda8d2a2b4a65aa2434a Signed-off-by: Arnaud J Le Hors <[email protected]> (cherry picked from commit fc736ec) Signed-off-by: Gari Singh <[email protected]> [FAB-5293] Log buffer overflow events This commit adds a logging event when a message is dropped due to overflow of the connection buffer to a remote peer In case of data blocks, the message that is printed is: 2017-07-16 09:12:55.442 EDT [gossip/comm] send -> DEBU 0c1 Buffer to 9.37.220.210:38346 overflown, dropping message GossipMessage: Channel: [121 97 99 111 118], nonce: 0, tag: CHAN_AND_ORG Block message: {Data: 107690 bytes, seq: 2771}, Envelope: 107714 bytes, Signature: 0 bytes Change-Id: Ibce121be0721a8c8a73f7805a051a28c213d64ba Signed-off-by: yacovm <[email protected]> (cherry picked from commit 28c8efd) Signed-off-by: Gari Singh <[email protected]> FAB-5331 Fix formatting of channel name Historically, gossip keeps the channel name in messages []byte instead of in *string*s. There is a String() method that formats complex messages It would be beneficial to print instead a string, for usability of log parsing. Change-Id: I71351fa324d6e1ca583588dab089a9e887996f7f Signed-off-by: yacovm <[email protected]> (cherry picked from commit ca167ce) Signed-off-by: Gari Singh <[email protected]> [FAB-5309] Set mod_policy for new channel policies The current configtxgen creates a channel creation tx with policies of Reader/Writer/Admin, where these each do not have a mod_policy set. This makes it difficult to change these policies, because the policy does not exist. It is possible through a roundabout way to change these policies (by removing and re-adding the policy with a different mod_policy), but this is certainly not a painless procedure and requires assistance of the orderer org. This CR updates to configtx processing to validate mod_policy values such that they are valid names. This means that the policy is not empty, and if the policy is path specified (ie using '/'), that each component of the path can be described as a valid config element (does not violate the config element naming rules). This CR additionally sets the mod_policy on the policies generated by configtxgen to "Admins". For users of the old v1.0.0 configtxgen, channel creation will fail with an error for v1.0.1 orderers. However, v1.0.1 configtxgen users will have their transaction appropriately consumed by v1.0.0 orderers. So in summary, upgrading configtxgen without fabric is okay. Upgrading fabric without configtxgen will cause breakage for new channel creation. An additional bug in the configtxgen output was discovered which was not setting a mod_policy on the anchor peers element. There was also an error in the bddtests which mimic the missing policies for the application level policies. Change-Id: Ic2bc120cfb6170f3e4e6cbeac5be145363a64861 Signed-off-by: Jason Yellick <[email protected]> (cherry picked from commit 3a2dd8e) Signed-off-by: Gari Singh <[email protected]> [FAB-5340] Respect new max message size on reconf Currently, the size filter only reads the max message size at construction time. This means that if this size is reconfigured, the filter will not work correctly until restart. This CR changes the sizefilter to re-read the size value each time. Change-Id: Ib61c74dd8031fc61e107dc10ab9eba9d8b6b866e Signed-off-by: Jason Yellick <[email protected]> (cherry picked from commit 36b08c7) Signed-off-by: Gari Singh <[email protected]> [FAB-5341] Solo should respect batchtimeout reconf Solo currently only reads the batch timeout at startup. This is a problem, because the batch timeout may be changed dynamically through a reconfiguration transaction. This CR simply causes solo to re-query the batch timeout from the config each time the timer is initialized, rather than once at creation time. Change-Id: I493fa4edddef40821336aa610ab242a3ffe97d26 Signed-off-by: Jason Yellick <[email protected]> (cherry picked from commit b817672) Signed-off-by: Gari Singh <[email protected]> [FAB-5339] Add missing nil check to extensions.go There is a missing nil check, that might make the code result in a null pointer panic. For reference: IsDataMsg is: func (m *GossipMessage) IsDataMsg() bool { return m.GetDataMsg() != nil } Change-Id: I5e20dee188e863b932312a7963f0ca73965c1457 Signed-off-by: yacovm <[email protected]> (cherry picked from commit 3cde835) Signed-off-by: Gari Singh <[email protected]> [FAB-5365] Fix bad error in peer CLI Deliver The peer CLI currently attempts to print the error status returned by the orderer's Deliver gRPC method. However, the log statement inappropriately uses the '%T' modifier, and prints the type of the status, not the actual status code inside it. Consequently, all deliver errors read the same uninformative error message: Got Status:*orderer.DeliverResponse_Status This CR fixes this log statement to include the status code instead, and additionally enhances the other error messages with pertitent information. Change-Id: I5a3e1dec574bfab178550cf67bc96a66f1896d5b Signed-off-by: Jason Yellick <[email protected]> (cherry picked from commit 22e1299) Signed-off-by: Gari Singh <[email protected]> [FAB-5342] spelling mistake in log: overflown Change-Id: Iadf71f9c8b34716a8d785767e73ab154a249df78 Signed-off-by: Scott Zwierzynski <[email protected]> (cherry picked from commit e3df726) Signed-off-by: Gari Singh <[email protected]> [FAB-5330] Prevent payload buffer overpopulation The state transfer module receives blocks from either the orderer or other peers, and puts them into a payload buffer for reordering so they would enter the ledger in-order. In some cases: - If the peer joined late and it receives blocks from peers starting from index i where the ledger is missing indices [j, i] for some j<i and the rate of block reception from peers is very fast - If the ledger is "stuck" (i.e file system full, etc.) and cannot advance This buffer would overpopulate. This commit addresses this, and adds a maximum distance constant that if the difference between the ledger height and the sequence of the block that is received is greater than this constant, the block is dropped. Change-Id: Ia1ba8966ea6d211c5d1b7ddd84a4fad34af797d4 Signed-off-by: yacovm <[email protected]> (cherry picked from commit ecda4c2) Signed-off-by: Gari Singh <[email protected]> FAB-5390 update chaintool URL update chaintool version and URL to download chaintool binary from nexus repository Change-Id: I871eea4106a6e0c59a3a9660d475026d6d2e5481 Signed-off-by: rameshthoomu <[email protected]> (cherry picked from commit 4223020) Signed-off-by: Gari Singh <[email protected]> [FAB-5407] Fix issue with pubkey pkcs11 CKA_PRIVATE According to PKCS11 standard: CKA_PRIVATE - CK_BBOOL - TRUE if object is a private object (vs. public object) (default FALSE) Changing this to FALSE for public key pkcs11 attribute struct fixed an issue we were having with our Thalas HSM: CKR_TEMPLATE_INCONSISTENT Commenting out this line also works (since there is no reason for it to be in the pubkey struct). Please also vendor in this change to other projects using it, such as Fabric-CA. Change-Id: I87d363a89e34c39f438492123166ed59eb515c44 Signed-off-by: gbolo <[email protected]> (cherry picked from commit c97a93a) Signed-off-by: Gari Singh <[email protected]> [FAB-5413] Add initial execution in retry process. Retry process should do an initial execution before starting retry time ticker, otherwise there is a delay for each retry, which accumulatively will cause slow start of kafka chain. Change-Id: I9c2aafeb73e366b72d94e317088548e4754ee8c4 Signed-off-by: Jay Guo <[email protected]> (cherry picked from commit 03afad8) Signed-off-by: Gari Singh <[email protected]> [FAB-5418] Add prereq for npm on Windows This change adds documentation on how to install the Visual Studio C++ tools necessary to build the Node.js native modules used in our samples. Change-Id: I78277d937bcb5e697620b122061c6050d2570448 Signed-off-by: Arnaud J Le Hors <[email protected]> (cherry picked from commit 7de02d2) Signed-off-by: Gari Singh <[email protected]> [FAB-5421] Add configtxlator to fabric-tools image Add configtxlator 2 fabric-tools docker image Change-Id: I36bd4677a393216e05d866aedac8a8f7206a96f3 Signed-off-by: rickr <[email protected]> (cherry picked from commit 210598f) Signed-off-by: Gari Singh <[email protected]> FAB-5422 make "Still have questions" prominent Change-Id: I46ab194a467dd5633b132f63e77505fe33f60a62 Signed-off-by: Christopher Ferris <[email protected]> (cherry picked from commit fbb84cc) Signed-off-by: Gari Singh <[email protected]> [FAB-5353]: Qualify sys. failure vs validation error Currently as stated in [FAB-5353], there is no clear separation during transaction validation during block commmit, between invalid transaction and some system failure which migh lead to inability to validate the transaction. For example db is down or file system is unavailable. This might lead to inconsistency of the state accross peers, therefore this commit takes care to distinguish between real case of invalid transaction versus system failure, later the error propagated down to the committer and forces peer to stop with panic, so admin will be able to take manual control and fix the problem therefore preventing peer state to diverge. Change-Id: I384e16d37e2f2b0fe144d504f566e0b744a5095c Signed-off-by: Artem Barger <[email protected]> (cherry picked from commit 0cf4c35) Signed-off-by: Gari Singh <[email protected]> FAB-5154 update prereqs to Docker 17.03.0-ce Change-Id: I34d23137eebfd9d5697061c1bf813b83543bcd05 Signed-off-by: Christopher Ferris <[email protected]> (cherry picked from commit 15c3bbe) Signed-off-by: Gari Singh <[email protected]> [FAB-5313] Leader election yield if deliver unavailable This commit makes the peer relinquish its leadership if it can't connect to the ordering service. Change-Id: I5fe679d5f23e539828fea4a9398c7dd4d9fd0f93 Signed-off-by: yacovm <[email protected]> (cherry picked from commit 6962ee3) Signed-off-by: Gari Singh <[email protected]> FAB-5422 fix syntax error Change-Id: Ied54138605d42b2c7687dc0f80a5a0eb3befca34 Signed-off-by: Christopher Ferris <[email protected]> (cherry picked from commit f80789d) Signed-off-by: Gari Singh <[email protected]> [FAB-5446] Fix orderer metadata local test The orderer metadata test only works if the development environment is setup to set common.Version, because if this is not set, the version is returned as "development version" and compared to the unset empty string. This is very annoying when trying to do development locally as this test always fails. This CR simply overrides the common.Version value for the purpose of testing this package. Change-Id: I6bc7a1de5a2730f4dd3fe17f1e5c7ff6bd81d4af Signed-off-by: Jason Yellick <[email protected]> (cherry picked from commit 7d59d5a) Signed-off-by: Gari Singh <[email protected]> [FAB-5391]Prevent concurrent invokes launching cc cont This CR prevents concurrent invokes from launching a chaincode container at the same time (e.g. during performance testing). The first invoke should succeed (and launch the container) while subsequent invokes should fail until the container has finished launching. Note: This does not change any behavior as subsequent invokes should have failed but it now sends a clear error message that the chaincode container is already launching. Change-Id: Ic1772a5f25dd0e4c34278e6e7bdb8507f16269c8 Signed-off-by: Will Lahti <[email protected]> (cherry picked from commit 2232d0e) Signed-off-by: Gari Singh <[email protected]> [FAB-5459] Recompute configmap instead of updating The configtx validation code works by transforming the ConfigGroup tree structure into a map, where each config element has a fully qualified key like: [Groups] /Channel/Application or [Policy] /Channel/Application/Readers This flattened structure makes it easier to check which elements have been modified, as the maps may simply be iterated over, rather than walking the config tree. After applying a config update, the current config map is copied, and the elements which were updated are overlayed onto the old config. This map is then converted back into a tree structure and serialized to be the new config tree. The current code adopts the updated config map as the current config map. However, this produces the bug described in 5459. Because the updated config map is the overlay of the new config onto the old config, the updated config may contain orphaned nodes which appear in the map, but which do not appear in the config tree. Consequently, when a subsequent update arrives, it is validated not only against the current config, but also against the orphaned nodes which are still in the updated config map. This CR changes the logic to discard the updated config map (which may contain this orphaned nodes) and instead has the config map recomputed every time a new config is adopted. This is more obviously deterministic and mimics the way a new ordering node would initialize the config after being restarted. Note: There is no accompanying test case with this. I had originally written a test case which demonstrated that nodes were orphaned in the updated config. However, this is expected and not a useful test case. Similarly, forcing the update config map to have updated nodes, then testing that that map is not adopted does not provide a valuable test case. So, instead of a test, this CR opts for some code comments and this lengthly commit explaining the change. Change-Id: Idc847cd5e237531e4a8b978f3465e30a909eb94f Signed-off-by: Jason Yellick <[email protected]> (cherry picked from commit 6eab9cf) Signed-off-by: Gari Singh <[email protected]> [FAB-5458] wrong type check in validator Validator is sending peer.TxValidationCode_INVALID_OTHER_REASON instead of peer.TxValidationCode_ENDORSEMENT_POLICY_FAILURE due to a wrong type check. Added UT to cover the switch and ensure correct tx flag been assigned. Change-Id: I7abe8828323780d527b24905f54d9a06e5f70b4e Signed-off-by: Srinivasan Muralidharan <[email protected]> Signed-off-by: Artem Barger <[email protected]> (cherry picked from commit e7b20bd) Signed-off-by: Gari Singh <[email protected]> FAB-5474 add links to godoc resources Change-Id: I13af4e7452cac28d0708ab4ba996af05f3fc8614 Signed-off-by: Christopher Ferris <[email protected]> (cherry picked from commit 65730c8) Signed-off-by: Gari Singh <[email protected]> [FAB-5329] Able to instantiate on a taken chaincode ID This fix appends a hash of the Docker image name to the image name to ensure a chaincode with different capitalization does not override the container of an already instantiated chaincode. This problem was introduced by the need to convert all Docker image names to lowercase. The Docker container name will retain its capitalization as that restriction only applies to the image name. Change-Id: I854f80b2f0e0269d9bbc60725f82b7e5a804b6fd Signed-off-by: Will Lahti <[email protected]> (cherry picked from commit 36e5140) Signed-off-by: Gari Singh <[email protected]> [FAB-5484] Fixes documentation for Windows users With this change the documentation instructs Windows users to work within the Docker Quickstart Terminal rather than Git Bash which does not work for Windows 10. This change also adds a note to the first curl command to address the recurrent problem with Windows users reporting failure due to using an outdated version of curl. Patch-set hyperledger#2: added a piece about installing npm grpc based on users reporting this was necessary. Change-Id: I95e7caeddc43e48f107955018aced112893ca540 Signed-off-by: Arnaud J Le Hors <[email protected]> (cherry picked from commit 1b6820a) Signed-off-by: Gari Singh <[email protected]> [FAB-5498] Missing RLock in onStatusChangeFactory There is a map access that isn't protected by a lock that was introduced in a recent commit. Change-Id: I31f884f76aaf0ba4a6b547c66830d8cd652f3c4e Signed-off-by: yacovm <[email protected]> (cherry picked from commit 201a814) Signed-off-by: Gari Singh <[email protected]> [FAB-5525] Fix configtx memory allocation bug The configtx code uses an 'append(...)' against a slice, and records the new slice address when computing the config map. Through sheer dumb luck, this code works for config groups which are only nested 2 levels deep, because the append call triggers a true new memory allocation. However, for config groups 3 levels deep (such as consortium groups), the append call actually re-uses the underlying memory for the slice. This causes the path to be corrupted internally for the consortium group items, and cause the wrong policy to be resolved when checking for a policy which has been specified relatively. This fix manually allocates new memory, copies it, and then appends the element. Change-Id: I0f4df619e006cdfebba60173156bda597d42a544 Signed-off-by: Jason Yellick <[email protected]> (cherry picked from commit b3c1430) Signed-off-by: Gari Singh <[email protected]> [FAB-5529] Missing check at endorser There is a missing check in the endorser code for a nil chaincodeID. Added a test. Change-Id: I10cb58ddb49cb9768aa2cdb9e9dc0b24808a5fbc Signed-off-by: yacovm <[email protected]> (cherry picked from commit 9b37c12) Signed-off-by: Gari Singh <[email protected]>
- Loading branch information
1 parent
d9c3202
commit 2cab745