Web Hosting Talk : Web Hosting Main Forums : Hosting Security and Technology : sda: asking for cache data failed

[norman1]

hello

First sorry for the long post

i order new server and after setup i see error given below in the Server Information

=======================================
sda: asking for cache data failed
sda: assuming drive cache: write through
sda: asking for cache data failed
sda: assuming drive cache: write through
sdb: asking for cache data failed
sdb: assuming drive cache: write through
sdb: asking for cache data failed
sdb: assuming drive cache: write through
==========================================

so i open ticket to the DC and the reply was

12 11:36
We are still looking into this issue.
Write caching should be enabled and working correctly by default on your hard drive.
I'm checking to make sure cPanel isn't misinterpreting the feedback from the sata drive.

===
12 11:40
I'm getting a slow response currently from the server, and after hooking up a console I see errors that all available memory has been addressed and is in-use.

I'm going to reboot the server now to get it responsive again.

===
12 14:13
Based on the information in WHM, it looks like the write-cache isn't reconized with the current driver set.

However, this is the most up-to-date set available for the controller on the sever. I can see that ncq is working correctly, and that the read/write times appear to be equivilent to other controllers with no issues.

As a work-around I have added blktool write-cache enabling settings in rc.local, and they appear to be setting the attributes correctly.

You are welcome to use any other drivers you wish to see if there is an increase in performance, but at this time everything appears to be operating normally.

Let us know if you have any performance issues with regard to drive speed, and we will investigate alternate avenues of resolution at that time.

===
12 15:09
Those values are pulled from dmesg on startup, and the work-around I have created executes after those values are pulled.

You should not have any issues (the drive "errors" aren't exactly errors, but the driver not being able to query the write cache on boot) and the performance of the drives appears to be up to par.

If you experience any slowdowns or high i/o wait in the future, let us know and we will investigate. However, you should not experience any issues.
==

then acunett send this mail to me
You should forward this to your DC techs and ask them to replace your HD:

root@host [/home/ahmed]# /sbin/hdparm -t /dev/sda

/dev/sda:
Timing buffered disk reads: 174 MB in 3.01 seconds = 57.80 MB/sec

******************************************************
root@host [/home/ahmed]# /sbin/hdparm -t /dev/sdb

/dev/sdb:
Timing buffered disk reads: 192 MB in 3.02 seconds = 63.50 MB/sec

Your HD seems to be slow for a sata drive.

the DC reply the development team work on this issue .

now i dont know if this Problem will caus more trouble in future ?
Would you advice what is the best way to solve this issue
Would you think that if thay change the server or change the hard drive this issue will be solved?

[EuroVPS/Director]

I don't think that 57-64MB/s would be considered slow at all for a SATA drive, I would not consider that a reason to replace a drive

What is the output of

# smartctl -a /dev/sda -d ata
# smartctl -a /dev/sdb -d ata

[norman1]

thanks for Re

the drive SATA II

==
root@host [/]# smartctl -a /dev/sda -d ata
bash: smartctl: command not found

[AcuNett]

At first glance that seems more like a driver/controller issue to me. Email the ticket ID to me and I"ll look over it. Those speeds are not slow, and I'd like to find out why someone on my side said that.