I was so bored with BCT down all day at work 
Ditto.
I was hitting the refresh button every 5 minutes to check if its back online.
Had to work all day at office because forum was offline
Topic: Recent downtime and data loss - page 4. (Read 6101 times)
Ditto.
I was hitting the refresh button every 5 minutes to check if its back online.
Had to work all day at office because forum was offline
Good work getting site back up.
I still prefer the old magnetic drives. I know they are slower, but they don't have all this cell wear problem. Also extra precautions need to be taken with SSDs if using as an encrypted drive, such as not using the TRIM command.
I still prefer the old magnetic drives. I know they are slower, but they don't have all this cell wear problem. Also extra precautions need to be taken with SSDs if using as an encrypted drive, such as not using the TRIM command.
That's what backups are for. If you don't have backups then you probably don't really care about that data anyway. Platter drives have their place, so do SSDs.
sure!!!
raid conf > 0, and RAID 1+0 too, is not to backup replace!!!
you need always to backup your data.
especially when it is possible to do hot backup like over DB.
Is the search feature is one of the problems? It disabled now 
legendary
Activity: 1862
Merit: 1011
Reverse engineer from time to time
I've said this before, SSDs are unreliable. If the thing cannot be used for 10 years without any wear, it's as useless as...incomparable.
Maybe I found out a bug/error. I sent a PM to theymos to correct it.
The error was 502:Bad Gateway Error (I got it while accessing a particular thread).
The error was 502:Bad Gateway Error (I got it while accessing a particular thread).
Good work getting site back up.
I still prefer the old magnetic drives. I know they are slower, but they don't have all this cell wear problem. Also extra precautions need to be taken with SSDs if using as an encrypted drive, such as not using the TRIM command.
I still prefer the old magnetic drives. I know they are slower, but they don't have all this cell wear problem. Also extra precautions need to be taken with SSDs if using as an encrypted drive, such as not using the TRIM command.
That's what backups are for. If you don't have backups then you probably don't really care about that data anyway. Platter drives have their place, so do SSDs.
Yesterday, internet was meaningless to me.
This is how the internet looked yesterday:
Well, well, well! We are seeing a new addiction that we hadn't realized existed before.
Yesterday, internet was meaningless to me.
Thank you theymos for getting the forum back online so quick.
Much appreciated
Much appreciated
I had no life for 12+ hours when your website went down. I'm glad its back up.
It really was rather blissful. Could you make it a regular thing?
Good work getting site back up.
I still prefer the old magnetic drives. I know they are slower, but they don't have all this cell wear problem. Also extra precautions need to be taken with SSDs if using as an encrypted drive, such as not using the TRIM command.
I still prefer the old magnetic drives. I know they are slower, but they don't have all this cell wear problem. Also extra precautions need to be taken with SSDs if using as an encrypted drive, such as not using the TRIM command.
SSDs are not reliable but then again, HDDs aren't reliable either.
I lost all my posts from yesterday.
I lost all my posts from yesterday.
not good at handling large amount of IO for server.
That's what I was thinking too. Same reason why you should defrag a SSD on a normal desktop.
Had several SSDs on server broken, I think SSDs are not good at handling large amount of IO for server. My traditional hard drive RAID never failed on the same server
I think I lost a few PM's, but nothing crucial.
Glad to see the forum is up and running. Thanks Theymos.
What SSD's are you running these on? Some SSD's are pure garbage in reliability, and OCZ is notorious for this. You can't really go wrong with Intel ones, although there was one series of Intel that had some random bricking.
Also you mentioned that these are fairly old SSDs. Just noting that SSD have a lifetime and do "expire", not by age but how much is written on there. I don't know much about server configuration, but if there are some intensive writing on the SSDs, especially MLC chips, they wear out much sooner.
Also you mentioned that these are fairly old SSDs. Just noting that SSD have a lifetime and do "expire", not by age but how much is written on there. I don't know much about server configuration, but if there are some intensive writing on the SSDs, especially MLC chips, they wear out much sooner.
That's what I expected to happen, but the RAID controller didn't notice that anything was wrong. I still need to figure out why.
IME the weak point of RAID is usually the controller: it's the non-redundant part of a redundant array :/
(be it because it's plain corrupting data, doing unnecessary I/O and wearing the disks... or just fails to report errors)
Jump to: