webmasterbeta
New Member
I ordered a managed dedicated server with a Raid-1 array configuration from a large and famous company based in Florida. I asked for a Raid-1 array so that I don't loose data and uptime, as there are some big guys hosted on this server. The sales guy said that if one HDD fails, they just replace it, and as expected, the Raid array is built in a few moments.
I'm the kind of guy who asks if the DC would handle a plane crash :uzi:
I do regular backups, but the web hosting company said I wouldn't loose data with such a configuration. Fine. Just to make sure, I e-mailed the support dept, and I asked how they would know one HDD would fail. Guess what: The guy told me one HDD was already broken and that the server was using only one HDD. I was literally amazed! :smash:
Is this normal behavior? For me, I didn't have to check every day to see through my obscure console if the Raid array was 100% operational. What if the 2nd HDD fails too? I would loose all my configuration and current data, right?If your second hard disk fails when the first one has already failed, you're losing all your data. That's why it's best to replace the broken HD immediatly when it's broken, otherwise you're losing everything.That's correct, most servers beep or flash lights when a disk fails (server grade hardware anyway). For example on Dell servers, the Dell logo turns red rather than the usual blue
.
Just to clarify, RAID does protect against failed disks, but not data corruption, so it's important to have backups too.
DanMost popular RAID cards come with software that your provider 'should' be using to determine the status of the RAID and it seems that your provider is. Aslong as your provider is pro-active on the RAID status and as soon as a failure is flagged they replace the failed hard-disk you should be in good hands.However I would ask your provider 'how long they have known that one disk has failed and why it hasn't been replaced?'This is why I asked, and their reply was "Your data should be safe"
Raid is not a backup. Get a back up solution also Raid is not backup, safest raid is raid 1+0 or raid10 with 4 drives. On the raid systems we have I have set up a cron job that runs a check and emails me twice a day the status of the raid so we can react pretty fast and replace any bad drives.I would deffinatly have the drive switched ASAP. Way better to be safe than sorry in this business!I couldn't have them replace it earlier because I have a national telco on this server and they have all their staff who access it all day long. It is now 23h30 and I just e-mailed the DC to switch the defectuous HDD. Downtime shouldn't be an issue.On some versions of Linux when you set up a RAID partition a process is run out of crontab to check the RAID status and send an email if a problem is reported.Your data is not safe if there is just one drive operational. Actually it isn't safe even if both are working, you should still have good backups.Yes, I have off-site backups thank God. Today they e-mailed me saying that actually the 2nd HDD wasn't defectuous, problem was that the Raid controller didn't have detected the 2nd driver properly, meaning that since the begining I didn't benefit from the Raid-1 array!
Oh well...What raid controller do you use? Nearly every raid controller comes with some kind of client interface, which allows you to check the array. Based on that interface you could also get some kind of automated script to check on the array every hour or so, so that it will inform you automatically in case of a failure.now at this point I would not know what to believe. maybe it was no issue at all !?you may know, there is hardware which allow hot-swapping - your benefit would be, if one harddrive fails it can be replaced without shutting downd the whole server. of course this hardware is not cheap but if you need them..auriance - BTW...i hope you dont get in trouble for copying rackspaces old designI have person experience with CHEAP and crappy RAID controlled and it was in a cheap Dell PE 400SC or CS. Its a promise controller (SATA) and it stores the raid array information on the Primary drive...lose that and you lose the array.. we had drive 2 fail first, then about a month ago drive 1 failed...replaced it...but the rai array wouldn't boot...we had to plug drive 2 directly into the MB...it runs..but we now have no raid...because the Promise control is a essentially a home office controller and need to remake the whole raid array from scratch now that it lost the setup on the main drive. On are expensive Power Edges we have raid 5, with hot swappable drive. That is sweet because you get a drive in 4 hours (we pay for that Gold support crap 
) and in a few hours after inserting you are back to normal...less the LARGE drive now days... I would recommend you check into that card...make sure drives are hotswappable...otherwise you maybe like me...buying a new server and making this junk server into a gaming rig 
Experiences like DA-Shane's are why I usually suggest software RAID1 rather than anything that relies on hardware. RAID5 can be a different story, since the controller becomes a smaller part of the overall cost as you are buying more drives and trying to have a lot of throughput on the system.The first thing you say is never say never. There are plenty of worst case senarios that would have all of your data destroyed, even something like a problem with corruption could take out your drives. Just remember your servers are only as good as the weaklest link so if you have something like a SOHO RAID controller do not expect enterprise level service.I would personally make sure the DC replaces that drive quickly then ride their *** until they give you a good explanation as to why it was not already changed out - most likely is that the hardware just does not alert them when a drive has failed like a higher end server would.I ordered a managed dedicated server with a Raid-1 array configuration from a large and famous company based in Florida. I asked for a Raid-1 array so that I don't loose data and uptime, as there are some big guys hosted on this server. The sales guy said that if one HDD fails, they just replace it, and as expected, the Raid array is built in a few moments.
I'm the kind of guy who asks if the DC would handle a plane crash :uzi:
I do regular backups, but the web hosting company said I wouldn't loose data with such a configuration. Fine. Just to make sure, I e-mailed the support dept, and I asked how they would know one HDD would fail. Guess what: The guy told me one HDD was already broken and that the server was using only one HDD. I was literally amazed! :smash:
Is this normal behavior? For me, I didn't have to check every day to see through my obscure console if the Raid array was 100% operational. What if the 2nd HDD fails too? I would loose all my configuration and current data, right?
Look into an rsync backup service like the one offered from bqinternet.com, this would definitely be a life-saver for you.
I'm the kind of guy who asks if the DC would handle a plane crash :uzi:
I do regular backups, but the web hosting company said I wouldn't loose data with such a configuration. Fine. Just to make sure, I e-mailed the support dept, and I asked how they would know one HDD would fail. Guess what: The guy told me one HDD was already broken and that the server was using only one HDD. I was literally amazed! :smash:
Is this normal behavior? For me, I didn't have to check every day to see through my obscure console if the Raid array was 100% operational. What if the 2nd HDD fails too? I would loose all my configuration and current data, right?If your second hard disk fails when the first one has already failed, you're losing all your data. That's why it's best to replace the broken HD immediatly when it's broken, otherwise you're losing everything.That's correct, most servers beep or flash lights when a disk fails (server grade hardware anyway). For example on Dell servers, the Dell logo turns red rather than the usual blue
.Just to clarify, RAID does protect against failed disks, but not data corruption, so it's important to have backups too.
DanMost popular RAID cards come with software that your provider 'should' be using to determine the status of the RAID and it seems that your provider is. Aslong as your provider is pro-active on the RAID status and as soon as a failure is flagged they replace the failed hard-disk you should be in good hands.However I would ask your provider 'how long they have known that one disk has failed and why it hasn't been replaced?'This is why I asked, and their reply was "Your data should be safe"
Raid is not a backup. Get a back up solution also Raid is not backup, safest raid is raid 1+0 or raid10 with 4 drives. On the raid systems we have I have set up a cron job that runs a check and emails me twice a day the status of the raid so we can react pretty fast and replace any bad drives.I would deffinatly have the drive switched ASAP. Way better to be safe than sorry in this business!I couldn't have them replace it earlier because I have a national telco on this server and they have all their staff who access it all day long. It is now 23h30 and I just e-mailed the DC to switch the defectuous HDD. Downtime shouldn't be an issue.On some versions of Linux when you set up a RAID partition a process is run out of crontab to check the RAID status and send an email if a problem is reported.Your data is not safe if there is just one drive operational. Actually it isn't safe even if both are working, you should still have good backups.Yes, I have off-site backups thank God. Today they e-mailed me saying that actually the 2nd HDD wasn't defectuous, problem was that the Raid controller didn't have detected the 2nd driver properly, meaning that since the begining I didn't benefit from the Raid-1 array! Oh well...What raid controller do you use? Nearly every raid controller comes with some kind of client interface, which allows you to check the array. Based on that interface you could also get some kind of automated script to check on the array every hour or so, so that it will inform you automatically in case of a failure.now at this point I would not know what to believe. maybe it was no issue at all !?you may know, there is hardware which allow hot-swapping - your benefit would be, if one harddrive fails it can be replaced without shutting downd the whole server. of course this hardware is not cheap but if you need them..auriance - BTW...i hope you dont get in trouble for copying rackspaces old design
I have person experience with CHEAP and crappy RAID controlled and it was in a cheap Dell PE 400SC or CS. Its a promise controller (SATA) and it stores the raid array information on the Primary drive...lose that and you lose the array.. we had drive 2 fail first, then about a month ago drive 1 failed...replaced it...but the rai array wouldn't boot...we had to plug drive 2 directly into the MB...it runs..but we now have no raid...because the Promise control is a essentially a home office controller and need to remake the whole raid array from scratch now that it lost the setup on the main drive. On are expensive Power Edges we have raid 5, with hot swappable drive. That is sweet because you get a drive in 4 hours (we pay for that Gold support crap ) and in a few hours after inserting you are back to normal...less the LARGE drive now days... I would recommend you check into that card...make sure drives are hotswappable...otherwise you maybe like me...buying a new server and making this junk server into a gaming rig Experiences like DA-Shane's are why I usually suggest software RAID1 rather than anything that relies on hardware. RAID5 can be a different story, since the controller becomes a smaller part of the overall cost as you are buying more drives and trying to have a lot of throughput on the system.The first thing you say is never say never. There are plenty of worst case senarios that would have all of your data destroyed, even something like a problem with corruption could take out your drives. Just remember your servers are only as good as the weaklest link so if you have something like a SOHO RAID controller do not expect enterprise level service.I would personally make sure the DC replaces that drive quickly then ride their *** until they give you a good explanation as to why it was not already changed out - most likely is that the hardware just does not alert them when a drive has failed like a higher end server would.I ordered a managed dedicated server with a Raid-1 array configuration from a large and famous company based in Florida. I asked for a Raid-1 array so that I don't loose data and uptime, as there are some big guys hosted on this server. The sales guy said that if one HDD fails, they just replace it, and as expected, the Raid array is built in a few moments. I'm the kind of guy who asks if the DC would handle a plane crash :uzi:
I do regular backups, but the web hosting company said I wouldn't loose data with such a configuration. Fine. Just to make sure, I e-mailed the support dept, and I asked how they would know one HDD would fail. Guess what: The guy told me one HDD was already broken and that the server was using only one HDD. I was literally amazed! :smash:
Is this normal behavior? For me, I didn't have to check every day to see through my obscure console if the Raid array was 100% operational. What if the 2nd HDD fails too? I would loose all my configuration and current data, right?
Look into an rsync backup service like the one offered from bqinternet.com, this would definitely be a life-saver for you.