top - 15:54:32 up 9 days, 49 min, 1 user, load average: 38.23, 30.61, 25.64
Tasks: 1777 total, 8 running, 1769 sleeping, 0 stopped, 0 zombie
%Cpu(s): 0.1 us, 75.2 sy, 0.0 ni, 22.7 id, 1.9 wa, 0.0 hi, 0.0 si, 0.0 st
KiB Mem : 32648604 total, 870392 free, 2642164 used, 29136048 buff/cache
KiB Swap: 16760828 total, 16552632 free, 208196 used. 18500956 avail Mem
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
1713 root 20 0 0 0 0 R 79.5 0.0 1439:34 txg_quiesce
1714 root 20 0 0 0 0 D 62.0 0.0 1310:11 txg_sync
我的服务器一次像这样卡住几分钟,没有进程可以使用 ZFS 文件系统进行 I/O。当它完成并恢复正常时,负载平均值从 1.0 以下开始上升到 80.0 以上。
我该怎么做才能阻止这种情况?
更新:
root@ca0:~# zpool iostat -v 1
capacity operations bandwidth
pool alloc free read write read write
---------- ----- ----- ----- ----- ----- -----
lxc 780G 1.00T 6.38K 0 8.51M 0
mirror 780G 1.00T 6.38K 0 8.51M 0
sda4 - - 3.25K 0 4.40M 0
sdb4 - - 3.13K 0 4.11M 0
---------- ----- ----- ----- ----- ----- -----
capacity operations bandwidth
pool alloc free read write read write
---------- ----- ----- ----- ----- ----- -----
lxc 780G 1.00T 4.69K 0 7.33M 0
mirror 780G 1.00T 4.69K 0 7.33M 0
sda4 - - 2.27K 0 3.42M 0
sdb4 - - 2.42K 0 3.91M 0
---------- ----- ----- ----- ----- ----- -----
capacity operations bandwidth
pool alloc free read write read write
---------- ----- ----- ----- ----- ----- -----
lxc 780G 1.00T 7.46K 0 10.3M 0
mirror 780G 1.00T 7.46K 0 10.3M 0
sda4 - - 3.72K 0 5.13M 0
sdb4 - - 3.74K 0 5.13M 0
---------- ----- ----- ----- ----- ----- -----
capacity operations bandwidth
pool alloc free read write read write
---------- ----- ----- ----- ----- ----- -----
lxc 780G 1.00T 7.37K 0 10.1M 0
mirror 780G 1.00T 7.37K 0 10.1M 0
sda4 - - 3.67K 0 5.02M 0
sdb4 - - 3.70K 0 5.03M 0
---------- ----- ----- ----- ----- ----- -----
^C
root@ca0:~# zpool status
pool: lxc
state: ONLINE
scan: scrub repaired 0 in 6h58m with 0 errors on Sun May 14 07:22:22 2017
config:
NAME STATE READ WRITE CKSUM
lxc ONLINE 0 0 0
mirror-0 ONLINE 0 0 0
sda4 ONLINE 0 0 0
sdb4 ONLINE 0 0 0
errors: No known data errors