Mac mini server with Mountain Lion and Promise Pegasus R2 harddrives

Question

Mac mini server with Mountain Lion and Promise Pegasus R2 harddrives

Hi, we have a Mac Mini server with Mountain Lion. The MacMini with accompanying external harddrives is configured as a file server and as a time machine server. We use Promise Pegasus storage units. These units consist of 6 * 2TB harddrives in a RAID5 setup. 2TB is wasted in the RAID5, so each unit has an effective capacity of 10TB.

Earlier we have had 1 Pegasus RAID Storage with Thunderbolt connected to the MacMini, and it has worked fairly well. Now we have purchased two new Promise Pegasus2 units, and connected these two as well to the MacMini. So now there are 3 Pegasus units in a thunderbolt daisychain.

Problem: copy jobs from the first Pegasus to any of the two new ones fail, and also make the system hang so that it needs to reboot.

When the copy job fails, a red led at one of the disks at the unit that is writing shows, indicating a failed drive. At the same time the activity led blinks, indicating that the same disk is spinning. The other disks show no red lights, and no spinning activity. So this inidicates a failed drive. The power-led on the unit that fails goes orange. However, on reboot the system shows no red or orange leds, and I am able to initate a new copy job. Which again will fail after 30 minutes or so. The disk that goes red varies, it is not the same one each time. And both new units has displayed this behavior (depending on which one is written to).

I have a copy script that copies several large directories, and the crash comes at varying stages in this script, indicating that the script itself is ok - once it even ran to completion. However - such an unstable system is not useful for us.

In the copy script I do:

cp -pPRX ...

without the X it does not work at all. So a problem with the Pegasus file system? It says Mac OS Extended (Journaled) on all drives - so it should be the same...

Any experiences of thoughts?

Mac mini, OS X Mountain Lion (10.8.5)

Posted on Mar 20, 2014 4:15 AM

Reply

Answer 1

Best reply

BDAqua

Level 10

238,997 points

Mar 20, 2014 4:10 PM in response to Snyltegjesten

Hi, no experience, but are you in the EU?

http://www.xlr8yourmac.com/archives/dec11/121911.html

Reply

Answer 2

Mar 21, 2014 6:13 AM in response to Snyltegjesten

I use Promise Pegasus2 R4 units on some Mac mini servers but only have one on each. Two are currently running OS X 10.8.5 and one is on 10.9.2, so far I have had no problems.

If you are mixing an original Pegasus R6 with a new Pegasus2 R6, then the original was Thunderbolt1 only and the newer ones are Thunderbolt2. As your Mac mini will be Thunderbolt1 anyway all three will be limited to Thunderbolt1 speed. I would however have if possible used all the same version.

What you should definately do is make sure the Promise software on the Mac mini is up-to-date and that the formware for each of your R6 units is also up-to-date. It would be safer to only have one connected at a time while you are updating the firmware.

If you run Disk Utility it will confirm whether each is formatted as HFS+ Journalled which is what you want and of course also using the GUID partition scheme.

Reply

Answer 3

Mar 21, 2014 8:34 AM in response to Snyltegjesten

Hi,

BDAqua - thanks for your reply. I do not have any indication that GSM900 radiation is the cause of the problems. However - I have not really tested for it either.

John Lockwood - thanks for helping out. As you point out - the speed is not optimal with my set-up. I achieved a copy speed of about 2.5 gigabits/s, that is 25% of the advertised 10 gigabits/s speed. However, this is of secondary concern, as long as my copy jobs crash.

The Promise Utility is updates. You are correct that this was important. Running an old Utility version with new driveunits made som of the information inaccessible. However, this was not the cause of my problems.

Firmware is also updated on all units.

This morning I returned the units to the Apple reseller where we bought them, for testing. Around noon they confimed that the filesystem on one of the units were broken - they were noe even able to mount the unit. They are going to format the disk. And so we will see.

Reply

Answer 4

Jul 18, 2014 3:57 AM in response to Snyltegjesten

Hi, long story short - I returned the Pegasus, got a new Pegasus and tested it. Also the second Pegasus had the error described above. So we returned that one as well, and bought a Synology DS1513+ instead.

I do not know for sure what the problem with the Pegaus'es was, but I suspect it is this:

http://features.techworld.com/storage/1019/what-is-time-limited-error-recovery/

Shortly summarized - the drives in the Pegasus RAID box were not designed for a RAID setup. So error handling in the drives, and in the RAID box, interfered with each other, making the box marking drives faulty when they were not - and the copy job stopped because of a TLER error (Time Limited Error Recovery). Some of the error handling on the individual drives should be turned off. We ended up buying a Synology DS1513+ instead, with WD Red drives. These drives are specifically designed for RAID setups. This machine completed the copy job without error. I think the Toshiba drives (DT01ACA200) that were in the Pegasus were not designed for a RAID setup. So it all boils down to - a design flaw at the Pegasus team?

Reply