作者归档：杨龙

mrq: ImportError: No module named coros

故障现象：

运行mrq示例（mrq-run tasks.Fetch url http://www.google.com），报错

[root@web4 test-mrq]# mrq-run tasks.Fetch url http://www.google.com
Traceback (most recent call last):
  File "/bin/mrq-run", line 9, in 
    load_entry_point('mrq==0.1.18', 'console_scripts', 'mrq-run')()
  File "/usr/lib/python2.7/site-packages/mrq/bin/mrq_run.py", line 33, in main
    set_current_config(cfg)
  File "/usr/lib/python2.7/site-packages/mrq/context.py", line 82, in set_current_config
    patch_import()
  File "/usr/lib/python2.7/site-packages/mrq/monkey.py", line 127, in patch_import
    import gevent.coros
  File "/usr/lib64/python2.7/site-packages/gevent/builtins.py", line 93, in __import__
    result = _import(*args, **kwargs)
ImportError: No module named coros

解决方法：

yum install python-gevent

./python/py_defines.h:38:20: fatal error: Python.h: No such file or directory

pip 安装mrp（pip install mrq）报错

./python/py_defines.h:38:20: fatal error: Python.h: No such file or directory

解决方法：

yum install python-devel

cp: cannot stat ‘modules/*’: No such file or directory

用php监控从库同步延迟个数

<?php
$mysqli = new mysqli('slave_host', 'root', 'passwd', 'dbname');

/*
 * This is the "official" OO way to do it,
 * BUT $connect_error was broken until PHP 5.2.9 and 5.3.0.
 */
if ($mysqli->connect_error) {
 die('Connect Error (' . $mysqli->connect_errno . ') ' . $mysqli->connect_error);
}

/*
 * Use this instead of $connect_error if you need to ensure
 * compatibility with PHP versions prior to 5.2.9 and 5.3.0.
 */
if (mysqli_connect_error()) {
 die('Connect Error (' . mysqli_connect_errno() . ') ' . mysqli_connect_error());
}

echo 'Success... ' . $mysqli->host_info . "\n";

/* Create table doesn't return a resultset */
/**
 * 
 * @var mysqli_result $result
 */
$result = $mysqli->query("show slave status");

if ($mysqli->errno) {
 printf("Error number: %s\n", $mysqli->errno);
 printf("Error message: %s\n", $mysqli->error);
}

if ($result !== false) {
 $row = $result->fetch_assoc();
 echo "master : {$row['Master_Log_File']}\n";
 echo "slave relay : {$row['Relay_Master_Log_File']}\n";
 echo " : " . ($row['Read_Master_Log_Pos'] - $row['Exec_Master_Log_Pos']);
}

$mysqli->close();

solr 突然卡死，重启失败，kill没反应，kill -9才杀死，重启后一会儿才好了

# service solr restart
Java not found, or an error was encountered when running java.
A working Java 8 JRE is required to run Solr!
Please install latest version of Java 8 or set JAVA_HOME properly.
Command that we tried: ‘java -version’, with response:
Error occurred during initialization of VM
java.lang.OutOfMemoryError: unable to create new native threadDebug information:
JAVA_HOME: N/A
Active Path:

出现这种问题的原因时solr内存配置不足，参考：

solr内存配置，solr时区配置

阿里云LVM 真·原地扩容加硬盘换硬盘释放硬盘

阿里云原地扩容方案：添加一个新的云盘替换旧的硬盘

第一步：扩容pv（物理卷）

创建物理卷（直接使用整个硬盘）

pvcreate /dev/vdc

将新物理卷（新硬盘）添加到卷组vg1

vgextend vg1 /dev/vdc

设置旧物理卷（旧硬盘）为不可分配

pvchange -x n /dev/vdb

将旧硬盘上的数据移走（热处理无损移动到其他硬盘确保其他硬盘空间充足）

pvmove /dev/vdb

可以将旧物理卷（旧硬盘）从vg1移除

vgreduce vg1 /dev/vdb

将旧物理卷（旧硬盘）从LVM移除

pvremove /dev/vdb

完成，可以拔除/dev/vdb了，此时可以使用pvs命令看到PFree列即为空闲空间

第二步：扩容lv（逻辑卷）

给lv +1GB容量

lvextend -L +1G /dev/vg1/lv1

使新加的容量生效（ext4）

resize2fs /dev/mapper/vg1-lv1

或者使新加的容量生效（xfs）

xfs_growfs /dev/mapper/vg1-lv1

LVM新加磁盘

新加磁盘为： /dev/vdc
LVM vg名称为： vg1

vgextend vg1 /dev/vdc

一次无损减小磁盘空间

[root@i ~]#
[root@i ~]# vgex
vgexport vgextend
[root@i ~]# vgex
vgexport vgextend
[root@i ~]# vgextend vg1 /dev/xvdb
Physical volume “/dev/xvdb” successfully created.
Volume group “vg1” successfully extended
[root@i ~]#
[root@i ~]#
[root@i ~]# ll
total 0
[root@i ~]# vg
vgcfgbackup vgchange vgconvert vgdisplay vgextend vgimportclone vgmknodes vgremove vgs vgsplit
vgcfgrestore vgck vgcreate vgexport vgimport vgmerge vgreduce vgrename vgscan
[root@i ~]# lv
lvchange lvcreate lvextend lvmchange lvmconfig lvmdump lvmpolld lvmsar lvremove lvresize lvscan
lvconvert lvdisplay lvm lvmconf lvmdiskscan lvmetad lvmsadc lvreduce lvrename lvs
[root@i ~]# pv
pvchange pvck pvcreate pvdisplay pvmove pvremove pvresize pvs pvscan
[root@i ~]# pvmove /dev/xvdb
No data to move for vg1
[root@i ~]# pvmove /dev/xvdi
Insufficient free space: 2292 extents needed, but only 1279 available
Unable to allocate mirror extents for pvmove0.
Failed to convert pvmove LV to mirrored
[root@i ~]#
[root@i ~]#
[root@i ~]# df
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/xvda1 41152832 2672832 36366516 7% /
devtmpfs 931496 0 931496 0% /dev
tmpfs 941156 0 941156 0% /dev/shm
tmpfs 941156 332 940824 1% /run
tmpfs 941156 0 941156 0% /sys/fs/cgroup
/dev/mapper/vg1-lvLog 3948224 16412 3708176 1% /mnt/log
/dev/mapper/vg1-lvData 5029504 453948 4297028 10% /mnt/data
tmpfs 188232 0 188232 0% /run/user/0
[root@i ~]#
[root@i ~]#
[root@i ~]# umount /mnt/log
[root@i ~]# df
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/xvda1 41152832 2672832 36366516 7% /
devtmpfs 931496 0 931496 0% /dev
tmpfs 941156 0 941156 0% /dev/shm
tmpfs 941156 332 940824 1% /run
tmpfs 941156 0 941156 0% /sys/fs/cgroup
/dev/mapper/vg1-lvData 5029504 453948 4297028 10% /mnt/data
tmpfs 188232 0 188232 0% /run/user/0
[root@i ~]# e2fsck -f /dev/vg
vg1/ vga_arbiter
[root@i ~]# e2fsck -f /dev/vg
vg1/ vga_arbiter
[root@i ~]# e2fsck -f /dev/vg1/lvLog
e2fsck 1.42.9 (28-Dec-2013)
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information
/dev/vg1/lvLog: 14/262144 files (0.0% non-contiguous), 53335/1036288 blocks
[root@i ~]# resize -f /dev/vg1/lvLog
resize2fs resizecons resizepart
[root@i ~]# resize2fs -f /dev/vg1/lvLog 500M
resize2fs 1.42.9 (28-Dec-2013)
Resizing the filesystem on /dev/vg1/lvLog to 128000 (4k) blocks.
The filesystem on /dev/vg1/lvLog is now 128000 blocks long.

[root@i ~]# lvresize -L 500M /dev/vg1/lvLog 500M
Physical Volume “500M” not found in Volume Group “vg1”.
[root@i ~]# lvresize -L 500M /dev/vg1/lvLog
WARNING: Reducing active logical volume to 500.00 MiB.
THIS MAY DESTROY YOUR DATA (filesystem etc.)
Do you really want to reduce vg1/lvLog? [y/n]: Y
Size of logical volume vg1/lvLog changed from 3.95 GiB (1012 extents) to 500.00 MiB (125 extents).
Logical volume vg1/lvLog successfully resized.
[root@i ~]# mount -a
[root@i ~]# df -lh
Filesystem Size Used Avail Use% Mounted on
/dev/xvda1 40G 2.6G 35G 7% /
devtmpfs 910M 0 910M 0% /dev
tmpfs 920M 0 920M 0% /dev/shm
tmpfs 920M 332K 919M 1% /run
tmpfs 920M 0 920M 0% /sys/fs/cgroup
/dev/mapper/vg1-lvData 4.8G 444M 4.1G 10% /mnt/data
tmpfs 184M 0 184M 0% /run/user/0
/dev/mapper/vg1-lvLog 364M 6.1M 323M 2% /mnt/log
[root@i ~]# ll
total 0
[root@i ~]# df
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/xvda1 41152832 2672836 36366512 7% /
devtmpfs 931496 0 931496 0% /dev
tmpfs 941156 0 941156 0% /dev/shm
tmpfs 941156 332 940824 1% /run
tmpfs 941156 0 941156 0% /sys/fs/cgroup
/dev/mapper/vg1-lvData 5029504 453948 4297028 10% /mnt/data
tmpfs 188232 0 188232 0% /run/user/0
/dev/mapper/vg1-lvLog 372680 6192 330652 2% /mnt/log
[root@i ~]# cd /mnt/log/
[root@i log]# ll
total 24
drwx—— 2 root root 16384 Nov 7 02:59 lost+found
drwxr-xr-x 2 redis redis 4096 Jan 13 20:10 mongodb
-rw-r–r– 1 root root 194 Mar 9 21:56 php_errors.log
[root@i log]#
[root@i log]#
[root@i log]#
[root@i log]# cd
[root@i ~]#
[root@i ~]#
[root@i ~]# df
Filesystem 1K-blocks Used Available Use% Mounted on
/dev/xvda1 41152832 2672836 36366512 7% /
devtmpfs 931496 0 931496 0% /dev
tmpfs 941156 0 941156 0% /dev/shm
tmpfs 941156 332 940824 1% /run
tmpfs 941156 0 941156 0% /sys/fs/cgroup
/dev/mapper/vg1-lvData 5029504 453948 4297028 10% /mnt/data
tmpfs 188232 0 188232 0% /run/user/0
/dev/mapper/vg1-lvLog 372680 6192 330652 2% /mnt/log
[root@i ~]#

API 的查询语言 GraphQL

GraphQL 是一个由Facebook提出的应用层查询语言. 使用 GraphQL, 你可以基于图模式定义你的后端. 然后客户端就可以请求所需要的数据集。

因此, 你不必因为客户端数据需求的变更而改变你的后端. 这解决了管理REST API中的最大的问题.

GraphQL同样能够让客户端程序高效地批量获取数据. 例如, 看一看下面这个GraphQL请求:

{
  latestPost {
    _id,
    title,
    content,
    author {
      name
    },
    comments {
      content,
      author {
        name
      }
    }
  }
}

这个 GraphQL 请求获取了一篇博客文章和对应评论与作者信息的数据. 下面是请求的返回结果:

{
  "data": {
    "latestPost": {
      "_id": "03390abb5570ce03ae524397d215713b",
      "title": "New Feature: Tracking Error Status with Kadira",
      "content": "Here is a common feedback we received from our users ...",
      "author": {
        "name": "Pahan Sarathchandra"
      },
      "comments": [
        {
          "content": "This is a very good blog post",
          "author": {
            "name": "Arunoda Susiripala"
          }
        },
        {
          "content": "Keep up the good work",
          "author": {
            "name": "Kasun Indi"
          }
        }
      ]
    }
  }
}

如果你使用的是REST的话，你需要调用多个REST API的请求才能获取这些信息。

GraphQL是一个规范

因此, 它可以用于任何平台或语言. 它有一个参考的实现 JavaScript, 由Facebook维护. 还有许多社区维护的实现有许多种语言。

介绍内容来自简书。

文章来源：https://www.oschina.net/p/graphql?fromerr=nRJiZA8n

Slave SQL: Error ‘Lock wait timeout exceeded; try restarting transaction’ on query.

Threads_running=54 exceeds its critical threshold 50

一般是负载过高，自动中断

加大这个阈值即可–critical-load Threads_running=400 加到400

pt-online-schema-change –user root
–password xxxxx
–host localhost
D=XXX,t=XXXXX
–execute
–alter “ADD UNIQUE INDEX invite_code (invite_code ASC)”
–nocheck-replication-filters
–charset utf8mb4
–max-load Threads_running=10
–critical-load Threads_running=400
–chunk-time=2.0
–no-check-alter