![[H]ard|Forum](/styles/hardforum/xenforo/logo_dark.png){kind=logo}
houkouonchi
RIP
- Joined
- Sep 14, 2008
- Messages
- 1,622
For quite a while I have praised areca controllers for their reliability compared to other brands. Unfortunately recently some things have happened which has caused me to have a change of heart...
We have had major issues with raid controller failures with both LSI and 3ware which is what prompted us to switch to areca. For the longest time we had no failures and just recently (2 months or so ago) various machines that have been running with no issues for 6 months or sometimes 1+ year are crapping out.
In linux the cards get arcmsr0: abort device command of scsi id = 0 lun = 1 and eventually will go to 'rejecting I/O to offline device'
I have seen this in the past caused by a bad drive getting timeouts or read errors and it was a matter of replacing the bad disk but recently its happened and there is nothing in the event log:
When a disk timeout occurred sometimes one of the disks would show a solid activity light but now a lot of times its like the controller is just crashing. Viewing the machine will show multiple random disks with hung disk activity lights. If you have a ethernet cable hooked up to it as soon as it crashes the web-interface will no longer be accessible. Here is a video I took of a server with a bad card and you can see the behavior:
http://www.youtube.com/watch?v=WNyjVWeehrk
This is right before I replaced the controller with a new one. So far replacing the controller with a new one has always fixed the issue suggesting its a problem with the hardware.
Most of the other brands of controllers we have had have failure rates between 15-25% (I am not joking) and until recently areca was <1% but now its getting up to about 5-7%. Its still much lower than the other brands but we have had problems with RMAing the cards. Areca checks out the cards and tell us there is nothing wrong with them =(. They said to run the newest driver and firmware and we were using the newest firmware but not driver. Even with the newest driver the problem persists and everything points to hardware issues.
We will likely be attempting to RMA again and hopefully get some new cards as it costs around $500 for each replacement card that we don't get an RMA for.
Anyway I just wanted to let people know my experience of the different brands as we buy in bulk. I have mentioned in the past areca has been much more robust/reliable than the other brands and now it appears they are more reliable but not 'much more' and they don't seem to be very good with RMA's even with a company that has over 500 of their controllers
Here is another vid I took of a bunch of dead raid controllers. We only ever had about 100 servers with 3ware (300 with LSI and 500+ areca) and you see we have a whole box of them and likely many others laying around.
http://www.youtube.com/watch?v=w8E2cu26jsI
We have had major issues with raid controller failures with both LSI and 3ware which is what prompted us to switch to areca. For the longest time we had no failures and just recently (2 months or so ago) various machines that have been running with no issues for 6 months or sometimes 1+ year are crapping out.
In linux the cards get arcmsr0: abort device command of scsi id = 0 lun = 1 and eventually will go to 'rejecting I/O to offline device'
I have seen this in the past caused by a bad drive getting timeouts or read errors and it was a matter of replacing the bad disk but recently its happened and there is nothing in the event log:
Code:
Date-Time Device Event Type Elapsed Time Errors
===============================================================================
2011-02-16 22:36:42 H/W MONITOR Raid Powered On
2011-02-16 12:31:49 H/W MONITOR Raid Powered On
2011-02-16 11:09:53 H/W MONITOR Power On With Battery Backup
2011-02-16 07:45:28 H/W MONITOR Raid Powered On
2011-02-15 14:41:47 H/W MONITOR Raid Powered On
2010-12-18 22:16:54 H/W MONITOR Raid Powered On
2010-12-18 22:07:45 H/W MONITOR Power On With Battery Backup
2010-12-18 21:43:01 H/W MONITOR Raid Powered On
2010-12-18 11:14:52 H/W MONITOR Raid Powered OnWhen a disk timeout occurred sometimes one of the disks would show a solid activity light but now a lot of times its like the controller is just crashing. Viewing the machine will show multiple random disks with hung disk activity lights. If you have a ethernet cable hooked up to it as soon as it crashes the web-interface will no longer be accessible. Here is a video I took of a server with a bad card and you can see the behavior:
http://www.youtube.com/watch?v=WNyjVWeehrk
This is right before I replaced the controller with a new one. So far replacing the controller with a new one has always fixed the issue suggesting its a problem with the hardware.
Most of the other brands of controllers we have had have failure rates between 15-25% (I am not joking) and until recently areca was <1% but now its getting up to about 5-7%. Its still much lower than the other brands but we have had problems with RMAing the cards. Areca checks out the cards and tell us there is nothing wrong with them =(. They said to run the newest driver and firmware and we were using the newest firmware but not driver. Even with the newest driver the problem persists and everything points to hardware issues.
We will likely be attempting to RMA again and hopefully get some new cards as it costs around $500 for each replacement card that we don't get an RMA for.
Anyway I just wanted to let people know my experience of the different brands as we buy in bulk. I have mentioned in the past areca has been much more robust/reliable than the other brands and now it appears they are more reliable but not 'much more' and they don't seem to be very good with RMA's even with a company that has over 500 of their controllers
Here is another vid I took of a bunch of dead raid controllers. We only ever had about 100 servers with 3ware (300 with LSI and 500+ areca) and you see we have a whole box of them and likely many others laying around.
http://www.youtube.com/watch?v=w8E2cu26jsI