Re: [pfSense] pfSense error, maybe hard drive?
On Thu, Mar 22, 2012 at 01:39, Jim Thompson j...@netgate.com wrote:
Hmm, No, close, but not really correct.
*all* flash will eventually fail if you write to it enough. It's physics.
I do not disagree of course. Fine with theory.
SLC NAND flash is typically rated at about 100k cycles, while MLC NAND flash
is typically rated at no more than 10k cycles. Via wear-leveling and
over-provisioning ('spare blocks') you can increase these numbers, but no
native flash device is rated in terms of millions of erase cycles.
You are talking about theory, the memory shell. I talk about the
actual flash disks.
There is a specific mechanism in these industrial flashes, doing
exactly this: When it finds an old memory shell refusing to be erased,
it re-allocates it (on the fly - transparently) to a healthy / not
used sector and marks it bad, much like a hard disk. Read their
documentation.
Now, imagine an 2G or 8G flash disk, containing only 50 - 150 MBytes
of code, like PfSense or any other small footprint OS or
application. It will practically never stop working. Many many
millions of erase cycles.
In real life, not theory. For 20 or 40 $ price.
Of course, if you anticipate many writes like data logging or
something, there are special techniques to overcome it. Like using RAM
and compressing/storing to flash every hour or so. All embedded
flavor OSs are doing things like this.
___
List mailing list
List@lists.pfsense.org
http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
On 21-3-2012 18:08, Adam Piasecki wrote: What hard drive is recommended for pfSense. Or can someone tell me what your running. Any ide or sata drive should do. If you really want a SSD drive I recommend the Intel 320 series SSD drives. These have a capacitor inside which means it will survive a power failure gracefully. We have 12 of those in a raid 6 (LSI Sas HBA, external enclosure), and another few in a raid 10 (Dell R610). We also have another 10 or so in various laptops and desktops and have had zero issues yet. We are planning to upgrade about 35 more desktops with the 120GB variant and 350 cash registers with the 80GB variant. We have about 70 Dell Optiplex 790 desktops which ship with the Samsung 830 series SSD drives which appear to work well too. For reference, I have a Corsair P256 (Samsung OEM SSD) which is still working well in my laptop. That's from the looks of it from july 2009 so it's now over 2 years old. Regards, Seth ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
On 21-3-2012 18:40, Jeppe Øland wrote: I deployed about a dozen Kingston 64G SSDs about a year and a half ago (in laptops and desktops) and I've seen about a quarter of them fail with different symptoms in each case. Garbage Totally agree. I have gone through 2 Kingston 4GB industrial SSDs so far - and it didn't take long either. They fail fast! (Now I'm using the 3rd one with an embedded install ... it seems to stay alive when nobody is writing to it). The dirty little secret from Kingston is that they do not manufacture anything themselves. The situation with SD/CF and microSD cards is horrific. You can easily end up with cards without proper production information indicating it's either from a test production runup or overtime production. Neither of which you want. http://www.bunniestudios.com/blog/?page_id=1022 The intel drives are a bit more coherent since they take a far different approach to manufacturing, they have used either their own 10 channel controller design (X25-M/320 series) or the Marvell controller (520 series). They coupled that with their own joint venture IMFT flash. That is a very tightly coupled process. Samsung does it very similar. The PB22J was a own design and memory, as was the 430 and 830 series. Which is probably the biggest reason for it's success with the large OEMs like Dell and Apple. Regards, Seth ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
On Mar 22, 2012, at 2:08, Dimitri Alexandris d.alexand...@gmail.com wrote:
On Thu, Mar 22, 2012 at 01:39, Jim Thompson j...@netgate.com wrote:
Hmm, No, close, but not really correct.
*all* flash will eventually fail if you write to it enough. It's physics.
I do not disagree of course. Fine with theory.
Theory here is reality.
SLC NAND flash is typically rated at about 100k cycles, while MLC NAND flash
is typically rated at no more than 10k cycles. Via wear-leveling and
over-provisioning ('spare blocks') you can increase these numbers, but no
native flash device is rated in terms of millions of erase cycles.
You are talking about theory, the memory shell. I talk about the
actual flash disks.
I believe I mentioned controller stunts to extend the lifetime of the flash.
There is a specific mechanism in these industrial flashes, doing
exactly this: When it finds an old memory shell refusing to be erased,
it re-allocates it (on the fly - transparently) to a healthy / not
used sector and marks it bad, much like a hard disk. Read their
documentation.
Yes, and I discussed this, but better than this is wear-leveling, which works
to avoid the issue, rather than reacting to failure. Combine this with some of
the advanced error correction, and you can greatly extend the lifetime of
(especially MLC-based) flash drives.
Apple the same tech to SLC-based drives, and their lifetime shoots up too.
So in the end, SLC will still win for endurance if your application does a lot
of writes.
The controller technology (over provisioning) you describe is at least 2
generations old. It works, but its nowhere near the state of the art.
Most CF cards can do the same thing now. (it's the source of the (harmless)
FreeBSD error with SanDisk CF cards, which report actual size, and then reserve
some percentage of sectors for this remapping.)
There are 32.5 million seconds or 8760 hours in a year. Writing once an hour
rather than once a second seems like an obvious way to reduce writes.
___
List mailing list
List@lists.pfsense.org
http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
On 3/22/2012 9:52 AM, Jim Thompson wrote: Yes, and I discussed this, but better than this is wear-leveling, which works to avoid the issue, rather than reacting to failure. Combine this with some of the advanced error correction, and you can greatly extend the lifetime of (especially MLC-based) flash drives. I have two questions, 1) Windows has TRIM support for ware-leveling. Does FreeBSD include this? Looking at the wiki page for TRIM (http://en.wikipedia.org/wiki/TRIM) it does not for 8.1, only for low level formatting. 2) If 8.1 does not support ware-leveling, would it be recommend that we not use SSD for pfSense until it does? Just trying to figure out if decent SSD (Not Kingston) would be recommend for pfSense. Thanks, Adam ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
For everyone, real world write tests (with synthetic writes), notice most drives able to write hundreds of TiB some approaching a PiB -- http://www.xtremesystems.org/forums/showthread.php?271063-SSD-Write-Endurance-25nm-Vs-34nm Adam - If you partition free space (under provision) the modern drives will wear level themselves quite well. Like Seth, we run a lot of Intel 320's, including our pfSense boxes with great success. We've been burned by cheap SSDs in the past though, for example we had a SanDisk unit that appeared to have a good 10x write-amplification in a pfSense box and died very prematurely. On the plus side pfSense continued to function properly without a hard disk indefinitely (though unmanageable). Mike On Thu, Mar 22, 2012 at 8:15 AM, Adam Piasecki apiase...@midatlanticbb.comwrote: On 3/22/2012 9:52 AM, Jim Thompson wrote: Yes, and I discussed this, but better than this is wear-leveling, which works to avoid the issue, rather than reacting to failure. Combine this with some of the advanced error correction, and you can greatly extend the lifetime of (especially MLC-based) flash drives. I have two questions, 1) Windows has TRIM support for ware-leveling. Does FreeBSD include this? Looking at the wiki page for TRIM (http://en.wikipedia.org/wiki/**TRIMhttp://en.wikipedia.org/wiki/TRIM) it does not for 8.1, only for low level formatting. 2) If 8.1 does not support ware-leveling, would it be recommend that we not use SSD for pfSense until it does? Just trying to figure out if decent SSD (Not Kingston) would be recommend for pfSense. Thanks, Adam __**_ List mailing list List@lists.pfsense.org http://lists.pfsense.org/**mailman/listinfo/listhttp://lists.pfsense.org/mailman/listinfo/list ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
On Thu, Mar 22, 2012 at 9:15 AM, Adam Piasecki apiase...@midatlanticbb.com wrote: 1) Windows has TRIM support for ware-leveling. Does FreeBSD include this? I can't speak to FreeBSD, but pfsense does not as of 2.0 2) If 8.1 does not support ware-leveling, would it be recommend that we not use SSD for pfSense until it does? Don't conflate TRIM with wear levelling. TRIM must be supported and enabled in the OS and the drive to work, while wear levelling is implemented by the drive independent of the OS. While wear levelling is designed to extend the life of solid-state storage, TRIM will actually shorten it (by way of increased write amplification) for the sake of preserving like-new performance. The down side of having TRIM support in pfsense is the gradual degradation of IO performance in the drive, which can be mitigated somewhat by garbage collection and spare area. See some of the many excellent SSD articles on anandtech.com for Anand's recommendations on which drives will maintain their performance better on non-TRIM platforms. db ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
On Mar 22, 2012, at 10:15 AM, Adam Piasecki wrote: On 3/22/2012 9:52 AM, Jim Thompson wrote: Yes, and I discussed this, but better than this is wear-leveling, which works to avoid the issue, rather than reacting to failure. Combine this with some of the advanced error correction, and you can greatly extend the lifetime of (especially MLC-based) flash drives. I have two questions, 1) Windows has TRIM support for ware-leveling. Does FreeBSD include this? Looking at the wiki page for TRIM (http://en.wikipedia.org/wiki/TRIM) it does not for 8.1, only for low level formatting. No, but FreeBSD 9.0 (which is to be the base for pfSense 2.1) does support TRIM for ffs. http://www.freebsd.org/releases/9.0R/relnotes-detailed.html#FS (answers from previous poster wrt TRIM .vs wear-leveling were also quite good) 2) If 8.1 does not support ware-leveling, would it be recommend that we not use SSD for pfSense until it does? Assuming you're asking about NAND-based SSDs... Just trying to figure out if decent SSD (Not Kingston) would be recommend for pfSense. Some of the better drive/controller combinations use superior forms of garbage collection, have a larger over provision of flash blocks, or are used on systems with a larger percentage of sequential writes .vs random writes. We're evaluating several SSDs here for inclusion on the pfSense systems we sell, but as this is a security appliance, and people tend to depend on it, we're stepping carefully. (This didn't actually answer your question, but I think Chris has already answered it.) Jim ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
On Wed, Mar 21, 2012 at 11:08 AM, Adam Piasecki apiase...@midatlanticbb.com wrote: What hard drive is recommended for pfSense. Or can someone tell me what your running. I use a Lexar Professional 2G and 4G compact flash with the embedded version in a couple of pfsenses. I deployed about a dozen Kingston 64G SSDs about a year and a half ago (in laptops and desktops) and I've seen about a quarter of them fail with different symptoms in each case. Garbage, and I'm done buying Kingston flash (I find their USB sticks unbearably slow too). By contrast I've used just about every Indilinx or Sandforce-based OCZ SSD in the last three years and yet to see a single one fail out of dozens deployed in various laptops, desktops and servers. db ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
I deployed about a dozen Kingston 64G SSDs about a year and a half ago (in laptops and desktops) and I've seen about a quarter of them fail with different symptoms in each case. Garbage Totally agree. I have gone through 2 Kingston 4GB industrial SSDs so far - and it didn't take long either. They fail fast! (Now I'm using the 3rd one with an embedded install ... it seems to stay alive when nobody is writing to it). On the OCZ Vertex drives, I would avoid Vertex1 ... they die pretty quick too if you write a lot. Vertex2 has been rock solid. Regards, -Jeppe ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
I always use the Intel 311 SSD's they are 20 GB SLC Drives priced very aggressive for a Single Level Cell drive, have yet to see one fail, and have shipped way over 100 of them for various embedded systems. Med venlig hilsen, Best regards Ulrik Lunddahl Sales Manager - Salgschef PROconsult Data A/S - Rugårdsvej 15 - 5000 Odense C Tel: +45 6311 - Tel dir: +45 63113341 - Mobil: +45 26363341 - Fax: +45 63113344 E-mail: u...@proconsult.dk - Web site: www.proconsult.dk -Oprindelig meddelelse- Fra: list-boun...@lists.pfsense.org [mailto:list-boun...@lists.pfsense.org] På vegne af Jeppe Øland Sendt: 21. marts 2012 18:40 Til: pfSense support and discussion Emne: Re: [pfSense] pfSense error, maybe hard drive? I deployed about a dozen Kingston 64G SSDs about a year and a half ago (in laptops and desktops) and I've seen about a quarter of them fail with different symptoms in each case. Garbage Totally agree. I have gone through 2 Kingston 4GB industrial SSDs so far - and it didn't take long either. They fail fast! (Now I'm using the 3rd one with an embedded install ... it seems to stay alive when nobody is writing to it). On the OCZ Vertex drives, I would avoid Vertex1 ... they die pretty quick too if you write a lot. Vertex2 has been rock solid. Regards, -Jeppe ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
On Wed, Mar 21, 2012 at 1:08 PM, Adam Piasecki apiase...@midatlanticbb.com wrote: I'm getting the following error when logging into the box. It's at the top of the page when presented with the username and password prompt. You can not go past the login page. pretty sure it's due to faulty hard drives. Indeed it is. We discussed this with the vendor you got them from at length, seems they got a bad batch of SSDs. Judging by recent experiences, I'd stay away from Kingston SSDs. ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
I'm getting the following error when logging into the box. It's at the top of the page when presented with the username and password prompt. You can not go past the login page. pretty sure it's due to faulty hard drives. Indeed it is. We discussed this with the vendor you got them from at length, seems they got a bad batch of SSDs. Judging by recent experiences, I'd stay away from Kingston SSDs. Are you saying you have discussed the issue with Kingston, and that they admitted problems? If so, are their problems resolved (ie. is it worth doing an RMA?) Regards, -Jeppe ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
On Wed, Mar 21, 2012 at 2:46 PM, Jeppe Øland jol...@gmail.com wrote: I'm getting the following error when logging into the box. It's at the top of the page when presented with the username and password prompt. You can not go past the login page. pretty sure it's due to faulty hard drives. Indeed it is. We discussed this with the vendor you got them from at length, seems they got a bad batch of SSDs. Judging by recent experiences, I'd stay away from Kingston SSDs. Are you saying you have discussed the issue with Kingston, and that they admitted problems? No, with the reseller that the OP bought the systems from. He was discussing with Kingston, we had to jump through some hoops to prove a hardware problem. I'm not sure where it went from there. ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
Re: [pfSense] pfSense error, maybe hard drive?
Normal commercial flash will eventually fail. It's not designed for this purpose. We use only industrial products which include error correction blocks and mechanism (transparent to the system), like: http://www.ieiworld.com/product_groups/industrial/detail_list.aspx?gid=1101cid=08141368770534315264 mainly IFM 4000+ / IFM 4400+ series: http://www.ieiworld.com/product_groups/industrial/content.aspx?gid=1101cid=08141368770534315264id=0A221362488516674830 We get tens of million cycles with not a single fail, in industrial environment (high temp. + vibration). On Wed, Mar 21, 2012 at 21:12, Chris Buechler c...@pfsense.org wrote: On Wed, Mar 21, 2012 at 2:46 PM, Jeppe Øland jol...@gmail.com wrote: I'm getting the following error when logging into the box. It's at the top of the page when presented with the username and password prompt. You can not go past the login page. pretty sure it's due to faulty hard drives. Indeed it is. We discussed this with the vendor you got them from at length, seems they got a bad batch of SSDs. Judging by recent experiences, I'd stay away from Kingston SSDs. Are you saying you have discussed the issue with Kingston, and that they admitted problems? No, with the reseller that the OP bought the systems from. He was discussing with Kingston, we had to jump through some hoops to prove a hardware problem. I'm not sure where it went from there. ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list ___ List mailing list List@lists.pfsense.org http://lists.pfsense.org/mailman/listinfo/list
