if the hdd get issue,what will the seller do ?

Portal Home > Knowledgebase > Articles Database > if the hdd get issue,what will the seller do ?

Posted by ttgt, 02-28-2017, 03:06 AM
Hi, if you order a raid dedicated server and you find certain one hdd has issue, you report the issue to the dedicated server provider to check and replace it. 1. do you think the seller should help you replace the hdd and void the server down asap ? 2. or he should shut down the server and take time to scan the hdd before replacing ? with my understand,it should be 1, because the server must be under production, if choose 2,and it is not the only way,it will stop the service to work. im not sure my understnading is correct ? or wrong ?
Posted by JiveNET, 02-28-2017, 03:15 AM
If this is a concern before leasing a machine (and lets be honest it should be), make sure that it has hot-swappable drives & HW raid and that their contract says that in the case of a disk failure they will replace the bad drive within some given period of time (without taking the whole machine down). I suppose for the most part you get what you pay for in terms of quality of hardware and support.
Posted by Srv24x7, 02-28-2017, 03:43 AM
Hi, You have a raid dedicated server, so it should be hardware RAID that is attached, so it will indicate if the disk is bad and once it is done, they can attach new one easily. The only issue comes in when it software RAID instead of hardware RAID because in software RAID you have to attach the disks backs in the array to rebuild it..
Posted by whmcsguru, 02-28-2017, 05:57 AM
HW raid is just a joke. I've never once in 15 years seen ir actually DO anything but cause problems and extensive downtime. No thank you. The best answer is proper backups. Take a good look at jetbackup from buycpanel. This thing does so much so right it's not even funny. If your hardware fails, you can restore from there
Posted by RackService, 02-28-2017, 08:54 AM
Indeed, you should not be using Hardware Raid. It's better to have your own backups (i.e) a storage server (another machine) where you store your backups to ensure those are always reachable even if your server is down. This to ensure you always have access to your files no matter what happens.
Posted by Chris-M, 02-28-2017, 12:18 PM
A RAID setup, either hardware of software based, is essential to prevent the failure of one drive from bringing down the whole system. I do not agree with the statement that "you should not be using hardware RAID" - it serves a purpose, but is absolutely not a substitute for backups. We would never deploy a production server without software or hardware RAID. As for the question asked by the OP, most datacenters will change a drive if sufficient evidence is provided by you to prove that the drive is failing. You should run a smartctl test and then provide the output of this to your datacenter - the output should show lots of errors and a FAIL result. That should be enough to confirm the drive is bad and you shouldn't need to do any offline testing. The next move would then be for the provider to change the drive which would involve only a few minutes of downtime while they perform the swap, unless you have hot-swap drives.
Posted by NortheBridge, 02-28-2017, 12:34 PM
Most people seem to thing RAID is somehow a backup strategy when it's not. It's there to not bring the entire system down when one (or more) drives fail which will happen eventually. One of the more recent servers we leased from our current go-to provider was provisioned with 16 brand new drives from HP but the RAID card was reporting a fault in one of the drives. Now everyone on the planet has access to SMART data and this drive reporting a fault hadn't been running for more than 30 hours (burn-in, load and hand-off). When a server is that new and one of the drives is reporting a fault on iLO it never hurts to investigate first. As literally the system was just being received and wouldn't be provisioned by us for another week we let them go about whatever they needed to do. In the end, they just replaced the drive and waited for the RAID to rebuild and after a few hours all health markers were green. All seamless and all painless.
Posted by madRoosterTony, 02-28-2017, 01:22 PM
I am glad to see more and more people start to speak out about RAID and its true uses, benefits and downfalls. As far as the OP goes, its all up to the companies policy and or if they have hot swap bay drives in the server. I can tell you why they are more common these days, not all companies use hot swap drive bays, so in this case the server would have to be taken down to replace the drive, so a quick scan might be helpful to verify what the smart data of the drive is reporting. Companies with hot swap bays should just swap out the drive if the smart data is reporting its failing.
Posted by arianna, 02-28-2017, 07:56 PM
in the past i had issues with HDDs when i purchased servers, always the seller cooperate to solve the problem, i sent the disk back to him and then he send me another one
Posted by ttgt, 02-28-2017, 08:49 PM
Hello, I have a server with very slow mdadm resync speed, I think it is not normal, And I find some hdd error from /var/log/messages. I ask dedicated server provider to replace the hdd and give them the error, They do not want to help me replace it directly, And they tell me some sector error is fine ,no need to replace, And they want to shutdown the server to check the hdd. Time later,they agree to hot swap the hdd the resync speed is normal now. So,I just wonder with normal way, when we really get issue and offer the error log, Server provider should replace the hdd asap ? Or shutdown the server to check the hdd ?
Posted by NortheBridge, 02-28-2017, 09:12 PM
The tolerance of "Uncorrectable Sectors" should be 0. At 1 the drive is disposed of or RMA'd or in the case of leased servers, replaced by the provider. If "Reallocated Sectors" is climbing exponentially that's also not a good sign.