Take the 2-minute tour ×
Server Fault is a question and answer site for professional system and network administrators. It's 100% free, no registration required.

On 29th October 2011, I built a RAID-5 array using 4 x 146.8GB Seagate SAS ST3146855SS drives running at 15k connected to a PowerEdge R515 with HP Smart Array P411 controller running Windows 2008 (so nothing particularly unusual).

I know that parity initialisation of a RAID-5 array can take some time but it's still running after 2.5 weeks which seems a little unusual.

I'd previously built another array on the same controller using 4 x 2TB SATA-2 drives and that did take a while to complete but a) I'm sure it was less than 2.5 weeks, b) that array was ~12 times bigger and c) during initialization, the percentrage slowly increased each day.

At the moment, the status display for this new 2nd array simply says "Parity Initialization Status: In Progress" and it's said that since the start. It's this lack of change on the status that worries me the most - feels like it's not actually doing anything.

Do you think something has gone wrong or am I being unpatient and for some reason, the status not increasing is normal? I kind of expected a much smaller array on faster drives (15k SAS versus 7.5k SATA-2) to build in a few days.

This is our primary SAN running StarWind so my "have a play" options are very limited. This 2nd array is currently in use for one small virtual disk so I could shut the target machine down, move the virtual disk to another drive and try rebuilding.

share|improve this question
    
So, just let me sum things up here. You put an HP controller in a Dell server and you didn't expect there to be problems? Yikes. –  MDMarra Nov 21 '11 at 14:38
    
37% accept rate. Did any of the answers help? –  ewwhite Nov 24 '11 at 12:49
    
@MarkM - no, I bought a controller that fitted the requirements we needed from eBuyer which happened to be made by HP in the same way it could have been made by LSI Logic. However, having read many HP posts on identical problems with HP servers and some SATA disks (inc. HP re-branded ones) and suspect it's not specifically some incompatability with just Dell. They've broken their own servers as well with the firmware upgrade –  Rob Nicholson Nov 24 '11 at 17:31
    
@ewwhite - thanks for your suggestions but as of writing, the problem isn't resolved and whilst your suggestion of upgrading the firmware is probably the right answer (and I will mark it as such next), it doesn't help for us as I don't want to risk upgrading the SAN (even though the SATA disks and not mission critical) and ending up with an unusable system. So I'm going to buy an LSI Logic card (which gives us 1GB cache and 6Gbit/s) and migrate the virtual disks across. Might then try upgrading it when not reliant on it! –  Rob Nicholson Nov 24 '11 at 17:37
    
I mentioned the cache module and flash/battery backup in my original post. I didn't realize your setup didn't have it. HP really shouldn't sell any of these controllers without cache modules and batteries. –  ewwhite Nov 24 '11 at 17:51
show 1 more comment

3 Answers

up vote 4 down vote accepted

Well, it's a little odd. I don't see many cases of mixing HP Smart Array controllers and Dell servers. Either way, the parity initialization doesn't begin until I/O is started on the new logical drive. May I ask how you're monitoring this? Via the HP Array Configuration Utility webpage? Perhaps the HP ACU command-line tool? If you have the latter installed, can you provide the output of:

ctrl all show config detail

We'd like to see that output to see if there's a potential issue with one of your disks.

From the HP Smart Array manual:

Background RAID creation 
When you create a RAID 1, RAID 5, or RAID 6 logical drive, the Smart Array controller must build the 
logical drive within the array and initialize the parity before enabling certain advanced performance 
techniques. Parity initialization takes several hours to complete. The time it takes depends on the size of the 
logical drive and the load on the controller. The Smart Array controller creates the logical drive, initializing 
the parity whenever the controller is not busy. While the controller creates the logical drive, you can access 
the storage volume which has full fault tolerance. 

Also, check the firmware on the Smart Array P411 controller. Do you have a cache module installed with a battery or flash backup? If not, you'll have other performance problems over time.

share|improve this answer
    
I'm using the HP Array Configuration Utility page so thanks for the heads up on the command line utility. I'll check that out and get back to you –  Rob Nicholson Nov 17 '11 at 16:10
    
Re: parity whenever the controller is not busy: hmm, I could see right now the controller been busy nearly all the time as the other array is been used a lot –  Rob Nicholson Nov 17 '11 at 16:12
1  
I don't think you can upload files to here and there is a small comment limit so I've put the output of the ctrl command here: mailbigfile.com/5e1532240f3984dcc0b2579c8165ba7e/…. That link will be valid for a few weeks –  Rob Nicholson Nov 17 '11 at 16:39
    
Upgrade the firmware! You're at v2.74. The current version is 5.12. Many bugs fixed. Download, install, reboot. Here's the link. h20000.www2.hp.com/bizsupport/TechSupport/… - changelog: h20000.www2.hp.com/bizsupport/TechSupport/… –  ewwhite Nov 17 '11 at 16:50
    
Thanks for the link, will schedule some downtime of the SAN to upgrade. Card was only bought a few months ago but I guess it could have been lying around in stock for a while. I moved the VM off today and have started the rebuild again. It now says "Queued" as the status but I agreed, I need to get the firmware updated ASAP –  Rob Nicholson Nov 19 '11 at 17:03
show 12 more comments

The likely-hood of a non-recoverable error in this day and age is extremely high. Might I suggest either a raid 1 or 10. Especially if this is holding anything important.

share|improve this answer
    
True. I would recommend against using RAID 5 for deployments these days for performance and reliability reasons. –  ewwhite Nov 16 '11 at 16:57
    
I agree for performance reasons but reliability depends upon how reliable. But for cost, RAID-5 is a good compromise. This array isn't mission critical - the internal disks in the R515 are RAID-10 and they hold the mission critical stuff –  Rob Nicholson Nov 17 '11 at 16:14
    
Actually RAID-5 is OK for small drives such as these. It's unreasonable for SATA drives, but for SAS or SSD drives it's OK. –  wazoox Nov 17 '11 at 17:08
    
Why unreasonable for SATA and okay for SAS? Parity rebuild time? The reason I'm asking is that we've got a large SATA (4 x 2TB) RAID-5 array connected to our Backup Exec server and speed is pretty pants but as this is "just" backup, the fact it takes two days for a full back isn't the end of the world (we also replicate to our USA site). Thought about putting 4 x 3TB drives in there and using RAID-10 to get a bit more speed –  Rob Nicholson Nov 19 '11 at 17:03
    
The size of SATA disks (2TB/3TB) versus these SAS drives (146GB) means you have MUCH more work to do, increasing your window for a double-drive failure. On top of that you're stressing your SATA drives (designed for 20% duty cycle) with near 100% duty cycle for a long time (days, possibly!) –  MikeyB Nov 19 '11 at 18:59
show 1 more comment

A disk firmware is available for DG072BABCE, and DG146BABCF drives : "This firmware prevents a rare condition that may occur during a WRITE SAME command sequence that may result in incorrect data being written to the hard drive. The WRITE SAME command may be used during RAID ARRAY parity initialization"

share|improve this answer
    
Do you have the URL where this information comes from. If so including it your answer will help. –  ChrisF Feb 3 '13 at 13:09
    
Can you reference where this firmware is available/downloadable? Also please reference the release notes if possible. Will make this answer much stronger. –  slm Feb 3 '13 at 13:09
    
This is moot. The OP is not using HP disks, so the firmware advisory does not apply. –  ewwhite Feb 3 '13 at 15:08
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.