Link to home
Create AccountLog in
Avatar of delphilenny
delphilenny

asked on

how can i identify a network proble / hard disk problem

i have a network that the performance is slow, i dont know if is because of the hard drive or network cable, there is a software that can help me determine what is failing?

in windows server 2003 i receive in a pop up windows that i have a corrupt file or unreadable please run chkdsk ( i have to delete and copy from backup or leaved alone, this message is a logical or the hard drive is failing?)

if i run the chkdsk when booting it take like 2 hours and i decided to esc. (deleting the file will take this away?)

Avatar of Britt Thompson
Britt Thompson
Flag of United States of America image

This sounds like a harddisk or disk controller problem. I'd make very sure you have a good backup. What type of controller/drive/system is this? HP, Dell, IBM... most of them have good utilities for their controllers that give you info about the drives.
Avatar of Dusty Thurman
Is the error reporting which file is the problem? If not, then we don't know how important that file is. It is highly recommended that you let the chkdsk finish. If it is taking a long time, the odds are it is fixing a lot of issues and/or you have a large amount of space to be checked. Are you using hardware raid? If so, what is the setup and what have to done to check the status of the raid?

As to isolating an issue between drives or network, there are lots of tools. The best are usually paid for and installed. Some are free and work well for most purposes. As far as free tools, the built-in performance monitor will suffice usually.

Start->Run and enter "perfmon". Once performance monitor is up, you will want to add counters. Under the category "Physical Disk", "Disk queue length" is a great one for harddrives. This should be nearly the same as your number of spindles (physical harddrives). Disk read/sec and Disk write/sec are good as well.

For the network, you might look at counters for Network Interface. Information about bytes/sec could be useful to see if you are getting values near your maximum.

From there, it will largely depend on your setup and what the server is used for. Terminal Services and other types of traffic on your server can be looked at with their own counters as well.
Avatar of delphilenny
delphilenny

ASKER

this is a raid mirror sata drives in a supermicro motherboard.
i have a total  backup .
should i delete the files that are giving me problems and copy again from backup. this will solve the problem or i should let over night the chkdsk utilyty?


I recommend letting chkdsk run first, it may find and fix problems you are not aware of yet.  Also I would look at the hard drive manufacture web to see if they have a utility that will let you check the drives health.
ASKER CERTIFIED SOLUTION
Avatar of Dusty Thurman
Dusty Thurman
Flag of United States of America image

Link to home
membership
Create an account to see this answer
Signing up is free. No credit card required.
Create Account
If chkdsk takes 2 hours to run the highly likely you have a lot of unreadable blocks.  It takes approx 30 seconds for a bad block to get remapped on consumer class SATA drives. If you do not have  SAS disks, or nearline SAS, or enterprise SATA running on Win2K, you have to get your head examined. Sorry, not going to be politically correct.

Enterprise SAS and enterprise SATA (in conjunction with redundant level RAID1,5,6,or 10) will typically recover unreadable blocks in 2-3 seconds, sometimes nearly zero seconds, depending on the RAID implementation.

ffff