迁移大list问题

[beyondblog]

大神们好！ ,在迁移过程中出现了一个错误 如下:

2017/02/15 18:25:55 [PANIC] parse command arguments failed
[error]: expect Array, but got <*redis.BulkBytes>
    2   /root/.gvm/pkgsets/go1.5/global/src/github.com/CodisLabs/redis-port/pkg/redis/resp.go:163
            github.com/CodisLabs/redis-port/pkg/redis.AsArray
    1   /root/.gvm/pkgsets/go1.5/global/src/github.com/CodisLabs/redis-port/pkg/redis/handler.go:88
            github.com/CodisLabs/redis-port/pkg/redis.ParseArgs
    0   /root/.gvm/pkgsets/go1.5/global/src/github.com/CodisLabs/redis-port/cmd/sync.go:258
            main.(*cmdSync).SyncCommand.func2
        ... ...
[stack]:
    0   /root/.gvm/pkgsets/go1.5/global/src/github.com/CodisLabs/redis-port/cmd/sync.go:259
            main.(*cmdSync).SyncCommand.func2
        ... ...

猜测可能是某个key的的value太大导致的

目前有个list 类型的key value 大约有几千万条 - - ...

我该如何解决这个问题呢？

[spinlock]

@beyondblog 这个我想过，需要改代码。现在的策略是直接 RDB restore 给 slave，单数据太大的话，肯定失败 (>512MB)，这种时候，看来需要手动的切成小的指令，比如 RPUSH 来操作。

[spinlock]

@beyondblog 不过从你的 log 上，你并不是这个问题。出错在于命令解析错误，不合法。

[spinlock]

@beyondblog 现在应该可以了，pull 最新的代码试试看。

f79bd50

[beyondblog]

@spinlock 非常感谢！ 稍后我试下! 😄

[luweijie007]

@spinlock 我也越到了一样的问题
2017/03/16 11:02:25 sync.go:235: [INFO] total=2908349195 - 2864928461 [ 98%] entry=5111118
2017/03/16 11:02:26 sync.go:235: [INFO] total=2908349195 - 2868982634 [ 98%] entry=5120844
2017/03/16 11:02:27 sync.go:235: [INFO] total=2908349195 - 2873281677 [ 98%] entry=5130978
2017/03/16 11:02:28 sync.go:235: [INFO] total=2908349195 - 2878549522 [ 98%] entry=5141108
2017/03/16 11:02:29 sync.go:235: [INFO] total=2908349195 - 2884614954 [ 99%] entry=5157670
2017/03/16 11:02:30 sync.go:235: [INFO] total=2908349195 - 2891919471 [ 99%] entry=5172628
2017/03/16 11:02:31 sync.go:235: [INFO] total=2908349195 - 2898386154 [ 99%] entry=5185882
2017/03/16 11:02:32 sync.go:235: [INFO] total=2908349195 - 2903861685 [ 99%] entry=5196923
2017/03/16 11:02:33 sync.go:235: [INFO] total=2908349195 - 2908300354 [ 99%] entry=5207640
2017/03/16 11:02:33 sync.go:235: [INFO] total=2908349195 - 2908349187 [ 99%] entry=5211041
2017/03/16 11:02:33 sync.go:237: [INFO] sync rdb done
2017/03/16 11:02:33 sync.go:259: [PANIC] parse command arguments failed
[error]: expect Array, but got <*redis.BulkBytes>
2 /usr/local/share/go/src/github.com/CodisLabs/redis-port/pkg/redis/resp.go:163
github.com/CodisLabs/redis-port/pkg/redis.AsArray
1 /usr/local/share/go/src/github.com/CodisLabs/redis-port/pkg/redis/handler.go:88
github.com/CodisLabs/redis-port/pkg/redis.ParseArgs
0 /usr/local/share/go/src/github.com/CodisLabs/redis-port/cmd/sync.go:258
main.(*cmdSync).SyncCommand.func2
已经是最新的代码了，是从一个redis服务导入数据到codis-proxy，每次都是到99%的时候，出现这个问题

[spinlock]

@luweijie007 这个很奇怪。。。

先说你已经完成 rdb 同步了，错误出现在追 log 的过程。但是 log 应该都是 BulkBytes Array 才对的。我不确定是不是 redis 某个版本改了协议还是我处理的有问题。

[luweijie007]

redis的版本和同步时候的日志如下，很正常，还有有用的信息我能查看的？
redis_version:3.0.6
3310:S 16 Mar 11:20:37.271 * Slave 10.15.107.143: asks for synchronization
3310:S 16 Mar 11:20:37.272 * Full resync requested by slave 10.15.107.143:
3310:S 16 Mar 11:20:37.272 * Starting BGSAVE for SYNC with target: disk
3310:S 16 Mar 11:20:37.279 * Background saving started by pid 6412
6412:C 16 Mar 11:21:24.306 * DB saved on disk
6412:C 16 Mar 11:21:24.307 * RDB: 6374 MB of memory used by copy-on-write
3310:S 16 Mar 11:21:24.415 * Background saving terminated with success
3310:S 16 Mar 11:30:00.668 * Synchronization with slave 10.15.107.143: succeeded
3310:S 16 Mar 11:30:08.991 # Connection with slave client id #20 lost.
是不是同步的数据太大6374 MB？
补充一下，redis服务的配置rdbchecksum 是关闭的，codis-port里面检查checksum的地方修改跳过检查，
这个有影响吗？

[spinlock]

不确定 如果怀疑的话 redis port 支持本地文件缓存功能可以加速 rdb 的获取。具体的你看一下 readme

…

On Thu, Mar 16, 2017 at 12:24 luweijie007 ***@***.***> wrote: redis的版本和同步时候的日志如下，很正常，还有有用的信息我能查看的？ redis_version:3.0.6 3310:S 16 Mar 11:20:37.271 * Slave 10.15.107.143: asks for synchronization 3310:S 16 Mar 11:20:37.272 * Full resync requested by slave 10.15.107.143: 3310:S 16 Mar 11:20:37.272 * Starting BGSAVE for SYNC with target: disk 3310:S 16 Mar 11:20:37.279 * Background saving started by pid 6412 6412:C 16 Mar 11:21:24.306 * DB saved on disk 6412:C 16 Mar 11:21:24.307 * RDB: 6374 MB of memory used by copy-on-write 3310:S 16 Mar 11:21:24.415 * Background saving terminated with success 3310:S 16 Mar 11:30:00.668 * Synchronization with slave 10.15.107.143: succeeded 3310:S 16 Mar 11:30:08.991 # Connection with slave client id #20 <#20> lost. 是不是同步不的数据太大6374 MB？ — You are receiving this because you modified the open/close state. Reply to this email directly, view it on GitHub <#33 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAsHpdYFnH6G2ojoVlSYURZANsWVH8f_ks5rmLlzgaJpZM4MBjWF> .

[luweijie007]

--sockfile=FILE [--filesize=SIZE，这2个参数吗？具体怎么设值，能贴下吗？