Ultra 2 Plus won't boot - support case #18604734

[mike1615]

I've been running a ReadyNAS Ultra 2 Plus for about four months now without any problems. A few days ago I upgraded it to the new 4.2.20 firmware, during which I think the system rebooted successfully as part of the firmware installation process. Last night it failed to restart after a scheduled reboot and I've been unable to get it back in service after several hours of working on it today. When I press the power button to turn it on, the fan starts running at high speed, the blue LED in the power button illuminates, and the little green activity LED illuminates. After about ten seconds the activity LED goes off but the blue power button LED stays on and the fan keeps running at high speed. I can hear the HDDs spin up normally, but the network interface hardware link/activity LEDs do not illuminate. It seems like the ReadyNAS is barely starting the boot process when it hangs, pressing the reset button has no effect, and RAIDar cannot find it on the network. I've swapped the HDDs with known good drives, but the ReadyNAS still won't boot.

I've seen a few other customers discussing similar issues online, so I'm starting to suspect that this is somehow related to the v4.2.20 firmware upgrade. Does anyone have any suggestions on ways to fix this problem or other things that I should try?

[chirpa]

If you boot it with no drives, and it doesn't show up in RAIDar saying 'no disks detected', there is something wrong with the chassis itself. You already have a support case, they will probably do an RMA.

[mike1615]

I don't think that booting without disks was any different (not visible in RAIDar), but I'll be sure to try that tomorrow morning just to be certain. It's really odd, four months of trouble free service then wham, it won't even boot up.

Thanks for your quick reply -- I'll update this thread as the problem is resolved; hopefully with enough details to make it an informative read...

[mdgm]

With electronics problems can emerge at any time. One reason why backups are recommended.

If an RMA is needed, I'd suggest putting a spare disk (must not be from your array) in the replacement unit, ensuring the firmware is up to date, then powering down and following the migration procedure: http://www.readynas.com/kb/faq/boot/how_do_i_migrate_disks_over_from_an_existing_readynas_to_another

[mike1615]

OK, I double checked it and even when turned on in diskless mode my ReadyNAS is not visible in RAIDar, so it looks like I'll be waiting to hear from Netgear regarding my support ticket.

I also noticed that after the activity LED turns off (about ten seconds after power-up), there's a small green LED inside the left, front side of the chassis that flashes rapidly for about three seconds, then it's off for about six seconds, then the cycle repeats. It's not visible on the front panel, only if I peer in through the air vent openings on the left side or when the chassis is in diskless mode it is also visible through the front grill. I haven't been able to find any documentation on it, though, and I'm reluctant to disassemble the chassis to investigate further since it's likely to be going back to the manufacturer for a warranty exchange.

[awbenway]

I just bought the same unit at Frys Sunday and it was working fine Sunday night with just 1 2TB disk. I too upgraded from 4.2.17 to 4.2.20. It rebooted fine, and I ran some tests copying about 80GB of Acronis images to it (on Gige was about 40% slower than my existing QNAP 219+). I then shut it down (via the GUI) to install drive #2, and it has never booted since then. The manual boot sequence works up to the point where you must press reset again to select the choice but it never does anything. I have tried cold starts (unplugged for 30 minutes) no-disk boots, new disk boot (pull first one), etc. to no avail. It will go back to Frys tomorrow. I just returned the Duo v2 (purchased last Monday) for almost exactly the same problem (no firmware update, but after maybe 5 reboots it was history - submitted support ticket, told to return it) to get this one. I guess I will have to order my second QNAP 219+ instead. If ReadyNAS units are this troublesome, I can't use them in my home office. I wanted to like it, but I've run out of patience.

[mdgm]

awbenway sounds like you've had a bad run. Most units work fine and don't have problems. I've got multiple ReadyNAS units and never had such a problem with any of them.

When you try to boot does the NAS show in RAIDar (http://www.readynas.com/downloads)? If so, what do you see under the Info (far right) column?

Can you ping the NAS?

If you can't see the NAS in RAIDar, what light pattern do you see?

Have you tried booting with just the old disk?

Also the ReadyNAS supports hot-adding new disks. No need to turn the NAS off before adding a disk. When you hot-add a disk it will be wiped and added into the array.

[awbenway]

Yep, seems like a run of bad luck lately.

The Ultra 2 is quite dead. With one disk in it, upon pressing Power the fan speeds up, the blue light comes on, the ACT green ligth flickers twice, the fans slow down - and then nothing. RAIDar can't find any units. It does not show up on my net. When you press Power again, it instantly shuts down (no OS lag, so none is alive).

I have tried booting with:
- 0 disks
- first disk with working image (initially) only
- new disk only

Always the same outcome. I have tried cold boots (power cable removed for 30+ minutes). The manual boot override also does not work. I can get to the boot menu and select what I want (I have tried Factory Reset and OS reinstall), but when I push Reset to select it (and go) nothing happens.

I am also concerned with how slow it was when it did run (one Hitachi 7200 rpm 2TB 64MB cache 6Gbps disk). On a test copying 16GB as thousands of small files, it was a a paltry 8-15 MB/s. On a few very large files (Acronis images, 15-23GB each), it reached 51 MB/s. On my QNAT 219P+, with 2 drives in RAID-1, on this same test (run immedialetly afterwards from the same host) I get 46 and 92 MB/s. The QNAP is $80 less ($300 at Newegg) than the Ultra 2 Plus ($380 here at Frys). I wanted to try a different unit (I am a computer engineer who is a senior performance and architecture staff member for a large storage vendor on internal lab testing - I like to play with new things!). And the QNAP has a size limitation problem of 2 x 1.5TB disks (and it is full now with a minimum depth of backups from my home office - my larger sets of enterprise array lab test data are about 16GB compressed, with 1000's of small files mixed with dozens of very big ones, what the Ultra 2 could only manage <15MB/s on). I will likely try a Synology 212+ next.

[StephenB]

On the speed issue - I found when I benchmarked my Pro that the number of drives in the raid array mattered.

The Bonnie plug-in measured the following [internal] speeds:

1 disk - 59.7 MB/s
2 disks - 110 MB/s
3 disks - 219 MB/s
4 disks - 319 MB/s

[awbenway]

Hi StephenB!

Actually, Bonnie (like IOzone) is a fairly useless tool. It measures your PC RAM and not the storage. You have to work with a target file that is at least 2x the size of your PC RAM to outsmart it. The best tools to use are Vdbench and IOmeter or SPECsfs (NAS specific). Also, note that the maximum lab rate for a single GigE connection for reads is about 108 MB/s (large block seq reads). For writes it is about 90-100. Anything higher is coming out of server buffers.

A Single RAID-0 disk is faster than a RAID-1 pair since all writes are duplexed, and the write completion ACK is not returned to the host until both are complete. Most entry level RAID controllers (what inexpensive NAS units and RAIKD cards for PCs are) do not read from both disks of a RAID-1 pair but only the primary. So reads are the same as one disk. For a 4-disk RAID-0 stripe, 4 disks should be (but many RAID controllers are poor quality and don't perform as expected) much faster than a single disk, but there would be no point in using RAID-0.

(Just tossing all this in since others will read this and may not know much about storage mechanisms.) When testing random small block performance, the metric to use is IOPS (not MB/s since that is for throughput - sequential loads). Random will never achieve the rates that large block (64KB - 8MB) seq can achieve due to the inefficiency of all the overhead to manage each tiny (2KB, 4KB, 8KB) I/O. Seq moves more data and the subsystem (should) perform prefetch upon seq detection of a stride pattern.

Anyway, on my Ultra 2 Plus problem (upgraded the firmware to 4.2.20 and after one successful reboot it is hosed), I heard back on yesterday's ticket (18617389) that I should try the boot menu and do a Factory reset - but the boot menu will not execute any choice made. So back the unit goes to Fry's tomorrow. The "level 2" techs at Netgear don't bother to read the details in the tickets it seems.

[StephenB]

Bonnie was running on the NAS (using Super-Poussin's add-on), so the test was not done over the network. At that point I did not want to enable SSH access, so I was sticking with the community tools.

The file size is > 2x the memory (2.1 GB file, with a standard PRO-6 with 1 GB ram total) . I tested 1, 2, 3, and 4 inserted disks, and the results went up about as I would expect. The NAS uses xraid2, so it is RAID-5 with 4 drives installed.

I agree that IOPS is the appropriate measure for random small block performance (though that point is perhaps more appropriate for the NFS tuning thread).

Are you exchanging your NAS for another ReadyNAS, or going with a different brand?

[awbenway]

Ah! I didn't know you could log into the ReadyNAS OS and run things. Hadn't really thought about. But if it's really a Unix box underneath, I should have expected that some expert Unix guys could figure that out (I'm not).

Support didn't answer my last response that I already related how the boot menu didn't work (could select a mode, but it wouldn't execute when you pressed reset again) and RAIDar and Frontview could not see it (because it never booted). I suspect that either the flash went bad, or the 4.2.20 update didn't install cleanly (but the Ultra did reboot cleanly and let me create a share and write a bunch of test files to it). It was the next shutdown cycle where it died.

I already have one QNAP 219P+ here, but it has a 1.5TB disk size limit (still does as far as I can tell). I may try the Synology 212+ which is also highly reviewed on SmallNetBuilder. But their name is too hard to remember most of the time... ;<)

[mdgm]

Does sound like it could be a flash problem. USB Boot Recovery of 4.2.20 could be worth a try: http://home.bott.ca/webserver/?p=166

Basically you download the zip file referred to in the link above but use the firmware image of 4.2.20 (http://www.readynas.com/RAIDiator_x86_4_2_20_Notes) not the older one in the zip file.

However there is a possible negative side effect of USB Boot Recovery which would require NetGear tech support to remotely login to the unit to fix.

Normally however if you're getting this problem with a new unit you'd want to simply return it as DOA to the reseller and get a replacement.

[readysecure1985]

I agree. give mdgms method a go.

[awbenway]

Thanks for the tips! My schedule is heavily overloaded with work, short selling my CA house, and moving to Oregon (trip #2 with U-Haul trailer tomorrow) - no time for experiments. I took the Ultra back to Frys this morning to return it within their 15 day window. I won't be back in CA until 6/2. So I have set all this aside for now. Netgear suport responded to my response to their response of yesterday today (see, I can emulate those marketing guys I -an engineer - have to deal with at work) and said to return it. Basic hardware failure on the board. Hmm, maybe there are some sort of "negative waves" at Frys that causes Netgear failures...... Two years or so ago they had a major run of bad power supply returns on all the Ready NAS units they sold - they pulled all ReadyNAS off the shelves for maybe a year.