• Content Count

  • Joined

  • Last visited

Everything posted by Quindor

  1. Quindor


    I have been doing some research into which drives might be compatible with my Adaptec 5805 raid adapter, or any hardware RAID adapter for that matter. In this search and my experience over the last years, RAID type disks do not exist for nothing. These disks are often picked from batches, better constructed, heavier and more robust to withstand 24H duty and thrashing around the disk. Also they often have tuned firmware to deliver the best possible results in a RAID and not so much in a desktop. And they sometimes have a different warranty period, compared to a desktop disk. Desktop disks have been traditionally a lot cheaper but not always safe to you use in a (hardware) RAID array. Sometimes problems would surface immediately, and sometimes only after a while. This lead me to find out what could be the main cause of this. In my opinion it has to do with TLER / CCTL. Everything else isn't really needed to make your RAID controllers understand the drive better. Let me try to explain : Which action should a HDD take in a situation where it has a hard time reading or writing something : Desktop disk - try unlimited times until success RAID disk - try 7 seconds max, then report back it failed to read the requested sector What do Controllers expect the disk to do : Desktop controller - Wait for the disk to respond, hangup and halt operations until success RAID controller - Wait max 8 seconds (often default value) then declare disk bad and evict/fail disk from raid set Ideal situation : RAID Controller with RAID disks. A disk develops a fault, tries for 7 seconds, then tells the RAID controller that it failed to read a sector. RAID controller acknowledges this and rebuilds sector from it's RAID, writes data to a different sector on the disks and marks the original sector as bad. RAID controller happy, Disk happy, no dropping disks! As you can see, mixing a desktop disk with a raid controller can produce unwanted results. Disks will be dropped, while nothing is really wrong with it, etc. In the end, it will become almost impossible to rebuild the RAID array anymore. Over time (when disks inevitably start to develop minor faults) this problem will surface and gradually become worse. It became known It first came to light that any disk can do either, and that it's just a value in the firmware when WD leaked a tool with which this could be set. They quickly disabled the functionality in their newer revisions of firmware and drives, to try to keep selling their much more expensive RAID editions (50% increase in price, nothing in cost, good deal for them). There is a good topic over at the HardOCP forum which work and which don't. WDTLER changed to setting in such a way that it would survive a reboot, actually altering the firmware itself. Breakthrough 1 But since the newest patches/SVN of smartmontools or smartctl, it's now possible to use SCT control to change the value the SCTECR (Error correction) is set to. Since this is done through SMART, any modern disk should support it. A big thank you goes out to r.gregory whom made this possible! Breakthrough 2 It's even now possible, using most modern RAID controllers, to use these commands on disks inside of a hardware RAID. And that's where it got interesting! Because if we can manually set the Error control values, we could ensure that at least this problem will not bite us again in the future! There is only a slight problem, and that is that the set value will be reset on every power cycle. Reboots mostly do NOT affect the value. Thus, running Windows or Linux, this is no problem, just create a script which sets it upon boot. Myself I run VMware ESX or ESXi on my servers and it posed a slight problem. My solution is a custom Fedora 12 USB bootstick, which boots in about 30 seconds, sets the value and automatically reboots. Pull the stick out during the reboot and the system boots from HDD with VMware on it. Problem fixed. So the only thing left now is to find out which disks are compatible with setting this command, and which are not. So, how do we test this..... Your controller does not matter (RAID controllers are bit more tricky and need a different approach though, see the smartmontools website). Windows users, download the windows installer. This is the newest 5.40 build. This gives you the smartctl executable, which we'll need. Users of other OS flavors will need to build their version from SVN, since the current binary versions do not yet have this functionality in it. Smartctl can be used for many other things besides this. It can tell you much about the life of your disk or say the current temperature for monitoring purposes, etc. etc. None of these tests below will touch your data. When you have that installed, and for example the disk you wish to check is your "d:" drive, execute the following : The Tests smartctl -l scterc d: If correct, this will give you the following feedback on a desktop disk Now we are going to try and change that value : smartctl -l scterc,70,70 d: If that works, you will see the following feedback : To put it back to original values again, either turn off the power of your system, or run "smartctl -l scterc,0,0 d:" Interesting tests are also if the value survives a reboot or even a power-cycle (from what I understand, it should not). Since I think this information is needed for anyone trying to build a home RAID array, we should keep a sort of database in this topic, maybe something like this: Brand Type Type2 Size RPM Revision Firmware Available Default Reboot Powercycle Samsung HD154UI F2EG 1.5TB 5400 - 1AG01118 Yes Disabled Stay Lost Samsung HD203WI F3EG 2.0TB 5400 - 1AN10002 Yes Disabled Stay Lost Samsung HD103UJ F1 1.0TB 7200 - 1A001110 Yes Disabled Stay Lost Brand Type Type2 Size RPM Revision Firmware Available Default Reboot Powercycle WD WD360GD 36GB 10000 00FNA0 35.06K35 No - - - WD WD10EADS Caviar Green 1.0TB 5400 00L5B1 01.01A01 Yes Disabled Stay Lost WD WD2500BJKT Scorpio Black 250GB 7200 ? 11.01A11 Yes Disabled Lost Lost Brand Type Type2 Size RPM Revision Firmware Available Default Reboot Powercycle Seagate ST31500541AS 1.5TB 5900 - CC32 Yes Disabled Stay Lost Seagate ST31500341AS 7200.11 1.5TB 7200 - CC1H Yes Disabled Stay Lost Seagate ST31000333AS 7200.11 1.0TB 7200 PCB rev A? CC1F Yes Disabled Stay Lost Seagate ST3250310NS ES.2 250GB 7200 PCB rev A? SN04 Yes Enabled (6s) N/A N/A Anyone willing to help and submit, much appreciated! Myself I would really like to know the results for the Samsung F3 2TB HD203WI EcoGreen disks! Update: Thank you all for submitting! Keep them coming!
  2. Quindor


    A word of warning, there seem to be problems with certain firmwares of the Samsung F4EG. Be cautious and read up about it before using smartctl on these drives!
  3. Quindor


    Yeah, that was my conclusion too. I directed him here from another thread, I believe he is working on it and will reply his results here! Has has never used SVN building before (which I had not also, before smartctl).
  4. Quindor

    PCIe eSATA Card Advice

    The VERY VERY cheap and dirty solution : http://www.addonics.com/products/host_controller/ad2sa6gpx1.asp . Not sure what you are wanting to do with it and if you would wish to use something like this in a business situation or not, but still. But sure to read the limitations. This card has shared bandwith (not non-blocking) and also should be used on a PCI-E 2.0 port if you wish maximum performance.
  5. Quindor


    Sorry for the very late reply. Not been around much lately, busy busy! Yes, WINDOWS software raid should not affect the ability of smartmontools. This is because you can still talk to the physical disk and thus you should be able to send commands to it directly. You might need to use their SCSI ID and not the drive letter, but that is a simple issue to solve. So if we are talking about say an intel ICH controller. No problem, software RAID or not. You need physical device access. The supported RAID controllers (like my adaptec) provide this.
  6. Quindor

    PCIe SSD Test Chamber

    ... I'm a bit at a loss here. What is a "test chamber" and what tests does it preform? See how the SSD holds up during extreme temperatures, etc.?? Heat should be faily easily to simulate yourself for instance. Please elaborate.
  7. Quindor


    I am actually the one using the Adaptec controller. Since I use VMware ESXi 4.0u1 on my server, I also have no way of running smartctl from the OS I am using. For this reason I use a bootable USB stick with Fedora installed. Fedora and their adaptec drivers are configured in such a way that next to your data arrays, they also provide a "SG" of your disks connected to the adaptec controller, a feature called "expose physical drives". With that, you can access the SMART data as if the disks are connected to a simple controller and also preform actions on them such as the Smartctl SCTERC commands. And off course read other stuff out of the disks. I also use this to run my Adaptec ASM software, since there is no version for VMware ESXi. Works great, if you do not have to powerdown your server too much. It takes about 5 mins in total to boot from the stick, set the desired commands and reboot into VMware ESXi. I know of no other way to get around this using windows, or even any other linux distribution for that matter. Normally I'm more prone to use Debian or Ubuntu, but I could not get the switch to work for me. Not a linux expert though. Hope this helps you, let us know!
  8. Quindor

    why so many large drive failures

    Ah, well, I have had dozens and dozens of disks go through my pc's/servers over the years. I have been running a storage server for years and thus upgrade once in a while and try to double the capacity when possible. This started with 20GB disks and has now moved to 2TB disks. So I have had 20GB's, 60GB, 120GB, 250GB, 500GB, 1TB, 1.5TB and now 2TB disks in my server (server itself has changed hardware too but in the most recents years I have been using adaptec controllers, so you could call that a constant. ). I've run from 4 disk RAID5's to 8 disk RAID5's. And have had some failed disks over the years.... treating them quite badly actually. I run the server 24Hr a day, for weeks on end, then shut it down, put it in the car, drive it to a lanparty, turn it on again for 48Hrs where the disks get trashed, turn it off again and put it back into it's spot where the disks spin again for a few weeks. Well, I can tell you, desktop disks (or any?) are not designed for that kind of usage. Oh, and my cars mostly have subwoofers in the back too, where mostly the server would stand also. Never had any effect on it that I could tell. But, since I have 3 years warranty mostly, I got by pretty well. Especially Maxtor was my friend with their advanced replacement program! :D Most disks would take 1 tot 1,5 or 2 years of this kind of punishment, in all honesty, the average lifetime you would expect from the disks in power on hours and reads, but normally over much larger time say 5 to 10 years. Anyway, in all my time with all my type of disks and array's and everything, I have never!!! lost as much data as I did this time. I have never had so many disks just fail, for no apparent reason. And now a days my disks are in nice holders in chieftec hotswap bays, etc. etc. So no, I am used to having a much better experience. Just like you are from what I get from your story. Sadly disappointed and I am hoping the replacement disks won't fail as horribly too!
  9. Quindor

    why so many large drive failures

    Just to add to this discussion a bit. 3 of the 7 Samsung EcoGreen F3EG 2TB disks have developed bad sectors this week and caused me a large amount of data loss. Quite a bad score if you ask me.
  10. Quindor


    Hmm, that's odd. That should work exactly as described. What did you set your controller to? It should be in AHCI mode for best result. Also, using your motherboard controllers RAID function will probably not give you the ability to use smartctl. But, using motherboard RAID is hardly or any better then using software RAID. Win2k3 does this fine, but consider Win2k8 R2... several great improvements! Also this will keep your array transportable between hardware, and also give you the ability to use smartctl! Win win situation I would think. Performance should be at least the same, you only forgo the ability to boot a RAID5. But create say a 50GB RAID1 (software raid, bootable) and then use the rest in RAID5. I've run like that before, works perfect! Let us know!
  11. Not sure what to awnser here, I can tell you though that 270MB/sec or even 200MB/sec is not attainable with a single drive. So I am not sure which specs you are looking at, but they do not sound realistic. About the reliability I am also not sure what to say. WD's are good, but Samsung's are not supposed to be bad. Currently having some trouble with my raid set, but I have other Samsungs which show no problems whatsoever.
  12. Nope. That was the old version. The new version has certified the X25-M's too and also you do not need to buy them from Adaptec anymore. This was announced at the Cebit and has since been built into the firmware from 2 or 3 weeks back. Also, more firmware updates should follow which will allow you to do write caching on an SSD as well. Personally I use my server for VMware ESXi purposes and I believe it could mean a great improvement for me! I have not tried the new firmware yet or have the SSD. But I believe it will be required to enter some kind of unlock code to enable the feature.
  13. Quindor

    Loving my new home server speeds

    Anymore specs or anything for us to go on here? Software RAID, Hardware RAID, SSD? Anything? Always good to share, but give us some specs and how you configured things, so others can learn from it also. Your speeds are a bit on the slow side I would say. A single disk should easily do 100MB/sec read and write at the beginning of the disk (Generation of 1TB disks or larger). So there might be something wrong, or there might not be. No clue....
  14. I too think this is a very interesting technology which will actually shape the storage to come in the next few years. I have an adaptec 5805 myself and am willing to pay for the SSD (consumer intel version), but don't want to buy the adaptec license for it. I runs a 7x5400 RPM 2TB array using VMware ESXi 4.0. Perfect enviroment I would say to see great SSD speed gains. Sadly, no key, no glory.
  15. What would be good is if you could do a media verification run on the disk (some controllers can do this in bios) or otherwise use the tools specific to the brand disk to test it if it's still good or not. For all of this you'll need to hook it up to a PC though.
  16. Ah, great to hear it's been fixed. I've only ever used the adaptec cables with the adaptec controllers. Will think twice about buying different ones now! Weird it did not work though.
  17. Very cool! But I sure hope they will make this option available for their existing controllers. I own a 5805 myself. From what I can see, the Q versions are just the same controllers, but with the software license for the SSD caching in it.
  18. Some left over metadata in on the HDD's could confuse the Adaptec controller. It should not though. If possible you could format one of the disks? Using a different controller or PC off-course. Other then that, are you running he latest firmware? That is also important! Let us know!
  19. Changed the fans on my Chieftec bays because the default fans are way way way too loud. Sure, they provide plenty of airflow, but I don't run 15K disks, but 5400rpm eco disks and some 7200rpm. Changed the fans, server is quiet again and disks (the 7200rpm's) only get slightly warm at best..... so cooling is still sufficient. Still very satisfied with the bays. Excellent quality. Please click the photo for the rest of the photos in the album.
  20. If everything goes right, my new 2TB EcoGreen Samsung disks will be in tomorrow. That will give me time to do some benchmarks tomorrow night, to stress the array, to benchmark the array and to test it's stability off course. I'm planning into taking it into production either Friday evening or Saturday. Planning to run the array using a Adaptec 5805 with 7 disks. The 8th disk position has been kept free to maybe use a MaxIQ module in the future. System specs (to test, array will be moved to server later on) : Asus Rampage II Gene Core i7 - 920 @ 4.0Ghz 6GB Memory Vertex SSD as boot disk, 1.5TB Samsung EcoGreen as data storage So, what would everyone like to see? I have my own tools I use, but I'd love to know what kind of tools you people use and which results you find interesting. You could call it a review, with which you can fill in the settings and benchmarks yourselves! Don't hesitate to respond, even simple suggestions are fine! Will fill the topic with screenshots and maybe a few photos later on! (when everything is in)
  21. Just wanted to report that since the build one of the Samsung 2TB Ecogreen's has failed on me. Crashed my raidset multiple times and alarms everywhere!! After being unable to find what was wrong I resorted to use the "full media scan" option inside of the adaptec controller. Scanning all the disks I found 1 disk that reported too many non-remappable errors. Bought a new drive, replaced the failed one, rebuilt the array and since then, no more problems! I'll send the drive for warranty and maybe use the drive I get back as a hotspare or for some other purpose. Further then that, everything is still working fine.
  22. Euhm, that will almost certainly NOT be the right choice, since it's a 5400RPM disk with aggressive power saving tendencies (can be tamed a bit). You'll want a 7200RPM disk. Disks aren't made to write multiple streams concurrently, or rather, they can't. They can do two actions, either write sequentially and achieve VERY high performance, or, seek. Seeking costs time and during a seek, nothing can be written. Writing multiple streams at the same time means that the disks will HAVE to seek to keep up with each sequential part on the disk. Seek times are related to platter density but mostly to spindle speed. So a 5400RPM disk take 3 times longer to do a seek them a 15.000 RPM disk will. Thus giving the 15.000RPM spindle a higher random I/O, while not necessarily giving it a higher MB/sec. This is also the reason a 300GB 15K RPM disk can have the same sustained sequential transfer rate as a 2.0TB 5400RPM disk. The 5400RPM disk needs 3x the platter density to negate the 15K RPM's spindle advantage. The 15K will always win in seeks though, because it can just reach a certain sector faster. Anyway, I'm getting off my point. Since you are writing multiple streams and your software probably is not intelligent enough to mix these on disk (mixing them or separating them both have their advantages either for writing or reading) I'm going to assume that they are written into separated files in a non-fragmented way. That means you'll have to handle a lot of seeks while writing this so your spindle speed is important. Cache will also greatly benefit you as NCQ will. Which brand to take, that I can't really tell you. More cache is good, so the 64MB variant you mentioned seems like a good start! Also multiple drives can greatly enhance how many streams your able to handle, again in relation to the above story.
  23. Man... I really get sick of these big companies doing stupid announcements like this. Tape backup or rather backup itself (also use Disk, Diskstage and Dedup) is my daily job. Tape technology will be on LTO7 in 4 to 5 years. This is per normal roadmap. LTO5 was released about a month ago and is shipping currently, and new LTO tech gets released every 2 to 2,5 years. This means that a tape costing about 40$ a piece will hold roughly 6.4TB natively. Adding 2:1 compression this comes down to about 12TB. Normally I'd calculate with 1,5:1 compression, but with LTO6 they will introduce larger compression buffers which should allow for a bit better compression ratio (not the 2,5:1 they envision). So, this guy could also say that in 4 to 5 years we will have SSD's disks of 12TB which will cost you only about 40$. Now that would be awesome! ... sadly... I think that to be complete fiction.... and not even close to what the truth is going to be. Sure, SSD's are great now and will only become better. They will push forward tech like 6G SAS and SATA as mechanical HDD's are not pushing these limits and won't be soon. Personally, I think 15K and partly 10K spindle disks will bite the dust sooner. You will get hybrid arrays which use 2TB Data disks (5400rpm or 7200rpm) with a 2 stage cache in front of it. Let's say 4GB memory cache per controller and 1TB of SSD cache as a sort of L2 cache. ZFS already can function in such a way. So, going with EMC standards what I think we'll have in a few years: 2x8GB memory cache (dual controller) 10TB SSD L2 cache 250TB SATA 7200 RPM Writes will go through SSD and cached read will also. This means your basically working on SSD but in the background it gets flushed to HDD. Actions can be streamlined (sort of advanced form of NCQ) and viola, the perfect, super fast, low response time solution. Doing Backup and Storage and also a bit of archiving. In almost all companies, looking at their array's only a fraction of their online data is actually used. So for a 250TB array, 10TB can hold almost all the data anyone in the company is working on. A lot of data gets stored and files grow bigger, but what is really being worked on.... doesn't change that much in size. Also, if you do this using a disk array. This technique can be applied block wise. A whole file does not need to be cached, just the blocks frequently used or being written. 7200RPM aren't slow, just not that good at seeking, so that needs to be cached. Anyway, I'm rambling now. Back to topic. I believe hybrid storage is the future. Dedup is getting important in the backup world, disk backups are fading fast and either used as a temporary stage for tape or being replaced for dedup storage. Tape is not dead, not by a longshot, neither SSD, Disk or Dedup hold the same advantages (yes, it also has disadvantages).
  24. Wow, very nice indeed!!! Having a 21TB server myself I especially can appreciate this project! Awesome little NAS system I'd say. I'd worry about cooling a bit though. The single fan could be enough, but it would produce noise, especially being so thin as he chose. Also, where is his intake and how does he ensure air passes between all the HDD's? Other then that, absolutely awesome indeed!
  25. Quindor

    Buffer Overflow on a SSD

    I truely have no clue what you are talking about.... Could you please explain your question? SSD's have no real Achillesheel that is very obvious. SDD's have advantages and Magnetic disks have advantages. So, please explain some more.