gian/cubefs
1
0
Fork 0
mirror of https://github.com/cubefs/cubefs.git synced 2026-08-02 02:00:56 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity
318 Commits 48 Branches 52 Tags 378 MiB
29805fe8f7
Commit Graph

318 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
awzhgw
29805fe8f7
Merge pull request #246 from chubaofs/dev 
merge from dev to master
 2019-09-11 22:14:14 +08:00
Liying Zhang
e4bd90129b Update CHANGELOG.md 
Release v1.3.0 - 2019/09/12
Signed-off-by: Liying Zhang <zly7755@163.com>
 2019-09-11 21:45:29 +08:00
awzhgw
8eaeabf5c1 Fix: datanode register hang bug 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-09-07 10:51:43 +08:00
awzhgw
2b2461598f Refactor: sync code from git.jd.com/chubaofs/chubaofs 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-09-06 21:06:46 +08:00
awzhgw
d5febd10da Fix: dataPartition disk error ,not recvoery raft log on new datanode 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-09-06 21:06:46 +08:00
zhuzhengyi1
dfca18fa1b fix: update export init for consul register 
Signed-off-by: zhuzhengyi1 <zhuzhengyi@jd.com>
 2019-09-05 18:22:54 +08:00
awzhgw
06a3ab3be2 Refactor: Increase the judgment condition of disk error,add 
syscall.EROFS

Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-09-05 08:40:21 +08:00
awzhgw
5e47ca00d3 Fix: if master only create DataPartitionCnt is 10,then cannnot mount 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-09-05 08:40:21 +08:00
Shuoran Liu
8ce0f64ad0 fix: err is shadowed in server main function 
Signed-off-by: Shuoran Liu <shuoranliu@gmail.com>
 2019-09-04 17:17:50 +08:00
zhuzhengyi1
c06e1dbaf7 fix: update export init for consul register 
Signed-off-by: zhuzhengyi1 <zhuzhengyi@jd.com>
 2019-09-04 17:17:23 +08:00
awzhgw
149d661e63 Refactor: metanode change deleteDentry or deleteInode api to log.LogDebugf 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-09-02 15:27:34 +08:00
awzhgw
f732ade2bc Refactor: log checkroration checkTime change to 1 second 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-09-02 15:27:34 +08:00
Shuoran Liu
4831443a56 fix: client gets stale file size if streamer is auto evicted 
If streamer is auto evicted, then extent client FileSize() will return a
generation of 0 and there is a chance for GetAttr system call to get
stale file size. So in such case, client should not trust inode cache
and send an iget request to metanode explicitly.

Signed-off-by: Shuoran Liu <shuoranliu@gmail.com>
 2019-09-02 15:27:22 +08:00
awzhgw
9c1e104822 Feature: support log module change loglevel on line 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-09-02 09:58:52 +08:00
Haifeng Liu
4b426360ee update readme 2019-08-30 22:51:46 +08:00
zhuhyc
3cf7c68015 fix: dataPartitionMap occurred error which is concurrent map iteration and map write 
Signed-off-by: zhuhyc <zzhniy.163.niy@163.com>
 2019-08-30 12:51:34 +08:00
zhuhyc
d69db16252 fix: the buf is not reset and the old data is repeatedly written to the ump warning log, resulting in the overall performance degradation of the cluster 
Signed-off-by: zhuhyc <zzhniy.163.niy@163.com>
 2019-08-30 12:51:34 +08:00
zhuhyc
6f0952fbd7 enhancement: the tickInterval and electionTick support reading from a configuration file 
Signed-off-by: zhuhyc <zzhniy.163.niy@163.com>
 2019-08-30 12:51:34 +08:00
Haifeng Liu
d3097ace5a
Merge pull request #228 from shuoranliu/daemonize-server 
feature: daemonize server
 2019-08-29 10:28:47 +08:00
Haifeng Liu
747c02f112
update readme 2019-08-28 18:31:33 +08:00
Shuoran Liu
df1a4d9cba refactor: keep mount point in error stataus when client is killed 
This commit will keep the mount point directory in error status when
client is killed, so the successive writes to this directory will not be
written to local disk. And the writer knows client is down as soon as
possible.

Signed-off-by: Shuoran Liu <shuoranliu@gmail.com>
 2019-08-28 18:21:01 +08:00
awzhgw
73355d2f1f Fix: Storage :when write tinyExtent,if offset!=e.datasize,return error 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-28 13:56:06 +08:00
awzhgw
b370e8d220 Fix: Datanode: when tinyExtentRepair auto repair,it has been 
repairTimeout

Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-28 13:56:06 +08:00
Shuoran Liu
ad20305923 feature: daemonize server 
Start server as a daemon, so no need to use nohup.

Signed-off-by: Shuoran Liu <shuoranliu@gmail.com>
 2019-08-26 16:05:01 +08:00
awzhgw
4dd740c435 Fix Datanode retain RaftLog: 
1.Datanode add log on recovery
2.Datanode Change raft log retain log to 10
3.delete unused file datanode/partition_metrics.go

Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-22 17:51:36 +08:00
Shuoran Liu
b15b78293b fix: fuse directIO read size can exceeds buffer size 
Since jacobasa fuse package uses a fixed-size memory pool for out
messages, it is important to tell fuse kernel module what the maximum
read size is. Otherwise, a large read will cause a panic.

However, max readahead size can only affect buffered read. The kernel
might still send a large directIO read in spite of the max readahead
size. We need to specify "max_read" option during mount.

Signed-off-by: Shuoran Liu <shuoranliu@gmail.com>
 2019-08-22 16:16:56 +08:00
Shuoran Liu
4f38ebad02 fix: treat ddelete not exist error as successful 
Signed-off-by: Shuoran Liu <shuoranliu@gmail.com>
 2019-08-22 08:46:50 +08:00
Shuoran Liu
5477452974 fix: metanode mistakenly delete empty dir inode 
If a directory is empty, the nlink of the dir inode is 2. And metanode
may mistakenly delete the inode upon receiving an iunlink request.

For example, if we are renaming a directory, the client will send ilink
and iunlink requests to the dir inode during the Rename process, and
this inode will be wrongly deleted after iunlink.

Signed-off-by: Shuoran Liu <shuoranliu@gmail.com>
 2019-08-21 18:31:32 +08:00
awzhgw
6f05a78e23 Refactor: Doc :delete warnLogDir config 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-21 17:51:52 +08:00
zhuhyc
97da53f065 fix: compatible with old heartbeat mode, old heartbeat mode does not report volname 
Signed-off-by: zhuhyc <zzhniy.163.niy@163.com>
 2019-08-21 10:39:34 +08:00
zhuhyc
d0c5e78b08 feature: the specified number of replica num is supported when creating vol 
Signed-off-by: zhuhyc <zzhniy.163.niy@163.com>
 2019-08-21 10:39:34 +08:00
zhuhyc
d6b118864b fix: The actual reduction in the number of replicas exceeds the expected reduction in the number of replicas 
Signed-off-by: zhuhyc <zzhniy.163.niy@163.com>
 2019-08-21 10:39:34 +08:00
Shuoran Liu
f63e5c657d fix: evict inode cache after successful deletion 
This commit evicts the inode cache after file or directory is deleted.

Signed-off-by: Shuoran Liu <shuoranliu@gmail.com>
 2019-08-17 20:16:31 +08:00
Shuoran Liu
7fff89b3c0 fix: exclude dir inode in the orphan list 
Directory inode should not be added to the orphan list, since it does
not need to be evicted.

Signed-off-by: Shuoran Liu <shuoranliu@gmail.com>
 2019-08-17 20:16:17 +08:00
awzhgw
dde79c0636 Refactor: add log on metanode delete extent 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
480c44af49 Refactor: DataNode: if not config raftDir,then not start server 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
6a9dbc369b Refactor: if not config warnLogDir,then donnot write umplog to disk 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
4a3ba7403f Fix: DataNode may be painc 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
f3297f91ec Refactor:Datanode delete unused func 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
4142d38ef9 Fix: Datanode deadlock on deletePartition 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
cc68ec3266 Refactor: when load dp,if dp status is normal,then start raft ,else wait snapshot has recover 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
eee4553dfe Refactor: docker: run ltptest print errorinfo 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
f9062c618d Refactor: Master create vol min default DataPartition set to 10 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
78bb09f858 Refactor: datanode start must start StartRaftLoggingSchedule func 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
07617e7f6f Refactor: datanode create dataPartition select disk function change: old: loop all disk,then create dp on this disk new: loop all disk select every disk weight(disk partition cnt *partitionSize /disk total size) 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
af066310af Refactor: docker metanode config.json change totalMem to 6GB 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
2701be38d8 enhancement: if create dp is normal,must start Raft else backend start raft 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
zhuhyc
abffb1712b enhancement: passing create data partition type to datanode 
Signed-off-by: zhuhyc <zzhniy.163.niy@163.com>
 2019-08-16 14:42:09 +08:00
awzhgw
67ce466ce1 Refactor: add ltptest log 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00
awzhgw
bcdfce2656 Refactor: DataNode: compatible old dataPartition Meta info 
Signed-off-by: awzhgw <guowl18702995996@gmail.com>
 2019-08-16 14:42:09 +08:00