Search results

Assuming you do a memtest prior to storing data, and you check the zpool status output for corruption regularly, I see no need for the emphasis on requiring ECC memory. Yes, ECC is a good idea for pretty much all computers. But despite some articles that suggest ZFS is very vulnerable to...

On BSD platforms there are the tunables. vfs.zfs.zil_disable vfs.zfs.zil_replay_disable However, disabling ZIL altogether is not recommended and I believe this tunable was removed in recent releases. It's not really required now that you can tune sync write behavior on specific filesystems...

Indeed. Is this thing actually for sale? It would be an excellent choice for a highend yet compact ZFS NAS server.

SLOG is nothing more than storing the ZIL on a dedicated drive. Without SLOG, your ZIL would be stored on the main pool instead. Sync=disabled means the sync writes will not be written to ZIL (intent log) first. Whether you are using SLOG (dedicated ZIL) or using regular ZIL, will not matter in...

You mean 'ad0-9' or 'ada0-9' I presume? I do recommend to flash your LSI controller in IT-mode firmware. IR will work, but may have side effects the IT doesn't have. The IR-mode firmware uses a different driver for the LSI2008 SAS controller chip. Whether you flash it or not, any disks...

If you do your research you will know Samsung is missing a few protections that other SSDs do have. Probably you would want to consider the M500 from Crucial as well. It is both cheaper than Samsungs offerings and offers more protections. It only offers lower sequential write speeds, but in many...

Today Sandforce isn't as bad as it used to be; all the OCZ consumers were used as guinea pig to iron out all the Sandforce bugs. But even with a stable controller, OCZ manages to screw up. For example by overclocking the controller for some extra benchmark results, but potentially higher RMA...

OCZ is nothing but a disaster for consumers, which get tricked by fake specifications and are awarded inferior products for the same price as honest and proper products. It's surprising how much second chances consumers award to OCZ, even though they are being ripped off multiple times. The...

A brand new SSD - even 25nm TLC one - will have about 10 years of retention when you start using it. Only when the flash degrades, will the retention drop significantly. When all the write cycles have been exhausted - on paper at least - you should still have 12 months of retention. So it...

Legacy RAID engines treat disks as either good or bad; nothing in between. A disk with terabytes of data but one tiny sector that cannot be read, is considered to be a failed disk from the viewpoint of the RAID engine. Of course, that is why you should not use legacy RAID, but instead migrate...

If by 'DMA errors' you mean UDMA CRC Error Count, then you replaced your drives for nothing; UDMA CRC Errors mean cabling errors and most likely are not the fault of the drive in question. @Topicstarter; i see no CrystalDiskInfo screenshot, just some pasted data. If it's right, then you've...

@omniscence, striping RAID can linearly increase IOps as queue depth grows, and for writes it requires less queue depth to scale. Of course, desktop appliations are limited in IOps and the SSD is much to fast for the system to really use all interleaving channels optimally.

Probably you are using Areca with non-TLER drives with uBER of 10^-14 or something like that. That is asking for trouble. Areca = legacy RAID. Your data is not gone however, once you clear the bad sectors on the disks you can reattach the detached disk and at least get a degraded array. I...

Not sure what the pricing is in your area, but here in Europe the M500 is a tad cheaper. But it packs extra protection in terms of power-safe capacitor array that can prevent retroactive datacorruption on SSDs. All SSDS based on NAND flash are vulnerable without power protection. People who...

TLER is no more than a workaround for faulty RAID engines. There is no reason a good RAID5 engine would need TLER. ZFS RAID-Z is one example; and the differences between RAID-Z and RAID5 are irrelevant to the harddrive. Upon reading an unreadable sector on a harddrive, the harddrive would...

The EVO doesn't have any power protection, and i consider it pretty expensive for a TLC SSD; it's not much cheaper than MLC at all. And the crappy software it gets shipped with, is horrible.

Ah, how nice to distort the names which ZFS uses by itself. :-( You must be among many others who were confused about this, jimphreak. But at least you know it's not your fault. :-)

A pool (or ZFS pool or zpool) is a storage volume that shares space with all filesystems contained. A pool contains at least one 'vdev' which is a virtual device. It is not that complicated. Let's try an example: You have one pool named 'tank' and it contains two vdevs. Both vdevs are like...

Why don't you try one of many ZFS NAS operating systems which can be configured with a web-interface. If you've never touched anything outside Windows, setting up Linux with command line stuff may be quite hard especially if the guide you're following doesn't work well for you. So I would...

The project I work on (ZFSguru) has a Handbrake service addon which you can install with two mouseclicks. Other services are also present like XBMC, Plex, PS3MediaServer, Transmission, SABnzbd+, etc. I do recommend a ZFS implementation based on the BSD operating system, which has the best...

Determine whether you need TLER. If your storage solution requires TLER, you need WD Red or comparable, which is TLER-enabled. If your storage solution is superior and does not require TLER, then get a simple Green or equivalent drive. For those solutions not requiring TLER, choosing WD Red...

Probably you are experiencing bad sectors, which causes your disks to drop out of the RAID. Do not use hardware RAID as it has the very same response to bad sectors: dropping disks from the RAID. I suggest you do not use RAID at all but rather single disks, or use a better solution like ZFS...

Well, I'm happy for you. But that does not invalidate my statement. In fact, it supports my statement. This is because your solution more than likely is not using regular consumer harddrives with 10^-14 uBER specification. Instead, you use something like 10^-15 or even 10^-16 harddrives which...

That is true. ZFS is the first reliable filesystem. Others will follow, but ZFS is the first reliable and usable solution. Take that as a fact, and move on. If you care about your data, you need to protect it properly. There is no solution which does this as well as ZFS. In particular, all...

16GiB RAM is on the low side for this class. Always buy the maximum capacity DRAM modules you can afford; buying 4GiB modules means a potential RAM upgrade is going to be very pricey. Buy 2 times 8GiB if you want to save money. As it looks now, you have an overkill excess of CPU power, but...

Metaluna, Solaris should handle GPT partitions fine, if you change the partition type to Solaris-usr. This can be done with ZFSguru via the web-interface or manually in Linux for ZFS-on-Linux. I would expect Linux to handle GPT partitioned disks better than Solaris, but I never tried this or...

Green drives are excellent for usage with ZFS. But I'm not at all convinced the harddrives are at fault. Could you please try another solution - preferably my own: ZFSguru - and boot it in a live environment. Import your pool and see how BSD ZFS identifies your pool. I will require the output...

If you do not need the controller to boot your operating system, you do not need the option ROM facility. You could flash your controlled with IT-mode firmware while not flashing the Option ROM at all. This will cause the controller not to display anything during POST (power-on-self-test; when...

You shouldn't need to re-add them. You simply reboot or in case of hot-swap this isn't even necessary. If the operating system resides on the pool that goes offline, you would be in trouble of course. The important thing is that ZFS just doesn't crap out like other RAID solutions do, kicking...

The biggest concern of ZFS on Linux is whether or not the implementation is mature enough. On Solaris and BSD platforms it took a long time before all the bugs and issues were ironed out and ZFS became a mature next-generation filesystem. On Linux the ZFS support appears to be improving, but I'm...

Intel 320 is the only reliable consumer-grade SSD right now. Soon the Crucial M500 can be added to that. These are the only SSDs for the consumer market that utilise power-safe capacitors. What olavgg said is unfortunately not true: This only works on mechanical harddrives. SSDs can die...

Samsung was already developing 1TB platters for 2,5" devices after they developed 1TB platters for 3,5" disks and demonstrated those some years ago before Seagate acquired them. 1TB platter is standard for 3,5" today but its introduction was stalled. The current top series of seagate/hitachi/wd...

Good news: http://wiki.illumos.org/display/illumos/Persistent+L2ARC

Any 'legacy' solution will mean you have no protection against bad sectors. That is a major problem IMO. So either you need proper backups of everything, or you should be willing to endure dataloss when bad sectors occur. Even worse, if they occur in metadata, a large portion of your filesystem...

Use GPT. You don't need to boot from it, so the limitations Microsoft Windows imposes on drives formatted with GPT do not apply.

The innovation has been out for some time. Ever since the Samsung M8 1TB notebook drive, there has been no new developments on this front. The maximum 2,5" platter capacity is still 500GB, despite Samsung's plans to develop 1TB platters for 2,5" drives. There is a WD notebook drive with 2TB...

One extra seek for one of 8 billion sectors isn't going to affect performance. Even with thousands of swapped sectors, this effect will not be noticeable. You need at least a million swapped sectors before this can actually impact performance in any meaningful degree. Next, virtually all...

Please consider his other thread as well: http://hardforum.com/showthread.php?t=1750553

The shops I order at are under a login, but prices have indeed risen since I last checked. I searched for some generic European price comparison sites and found this: HP SAS Expander http://www.heise.de/preisvergleich/hp-smart-array-sas-expander-card-468406-b21-a514339.html 200 EUR ~= $260...

Probably because you are not thinking in terms like a ZFS box, but rather conventional RAID done by the controllers firmware. In that case, a single RAID controller with SAS expander would be more logical as I already stated in my earlier message. For ZFS systems, you do not want RAID...