pueblosnet

Active Member
Dec 23, 2003
38
0
156
Until now I used servers with a maximum of 2TB of hard disk and used RAID 1 for software without problems. It took about 1 day to synchronize. Now I am faced with the problem that with the 4TB disk it will take about 1 week and when it does, it will have to start again.
I have increased the writing speed to the maximum in the files:

/proc/sys/dev/raid/speed_limit_min
/proc/sys/dev/raid/speed_limit_max

And that has somewhat improved the timing, but not much. This is the result of:

[[email protected] ~]# cat /proc/mdstat
Personalities : [raid1]
md2 : active raid1 sda3[0] sdb3[1]
3872806720 blocks super 1.2 [2/2] [UU]
[=======>.............] check = 38.3% (1486756992/3872806720) finish=34929.4min speed=1138K/sec
bitmap: 20/29 pages [80KB], 65536KB chunk
md1 : active raid1 sda2[0] sdb2[1]
523712 blocks super 1.2 [2/2] [UU]
resync=DELAYED
md0 : active raid1 sda1[0] sdb1[1]
33521664 blocks super 1.2 [2/2] [UU]
resync=DELAYED
unused devices: <none>

That could be happening? I appreciate any suggestions or ideas on what else to look at, thanks.
 

sparek-3

Well-Known Member
Aug 10, 2002
1,929
178
343
cPanel Access Level
Root Administrator
Assuming since these are such large drives, that these are spinning SATA drives. Yea, it's going to take a while if you are using 4TB disks, that's one of the downsides of using such large disks. Spreading this out among multiple servers might be a better solution. Or perhaps 4 2TB drives in a RAID10.

But 1138K/s is very slow.

The raid check is going to depend on how much disk I/O is free to be used on the drives. So if you have something disk intensive (like a backup, cpanellogd forming statistics, or any other disk intensive tasks) it's going to slow down the process.

You might consider switching the check out to monthly. I always thought weekly was overkill with this.
 

cPanelLauren

Forums Analyst II
Staff member
Nov 14, 2017
7,975
640
263
Houston
cPanel Access Level
DataCenter Provider
Hello @pueblosnet

As indicated by @sparek-3 for that size disk it will take a considerable amount of time and the check depends on the amount of resources (I/O ) you have available to run the check. With the check running at 1138K/sec I would assume you've got some disk intensive tasks occurring. Other than the suggestions that have been made already (which I the optimal solutions personally) the only other thing is you may want to limit the amount of these tasks that are occurring on the server during the synchronization though this may be an unrealistic suggestion depending on how much attention you want to be devoting to this.