Decoding MPT_SAS drives in Nexenta/Illumos

Another group on campus had SAS resets on a drive, but the drive never failed. All we ever got was something like this in dmesg reports:


May  9 07:51:50 hostname scsi: [ID 365881 kern.info] /pci@0,0/pci8086,d138@3/pci1000,3010@0 (mpt_sas0):
May  9 07:51:50 hostname  Log info 0x31080000 received for target 9.
May  9 07:51:50 hostname  scsi_status=0x0, ioc_status=0x804b, scsi_state=0x0

We have a target address, but a zpool status or equivalent never gives us that target number, just a long string like "c0t50014EE057D81DE7d0". How to find the drive to off-line and replace it?

For future reference, you'll need to follow this link:

https://www.meteo.unican.es/trac/meteo/blog/SolarisSATADeviceName

You'll also need the lsiutil, which I is part of the LSIUtil Kit 1.63. For reference: One valid download link --- Its next to impossible to find this on LSI's site and the Oracle references are all down now it seems.

Joyent Anniversary: What's Next

It was just over a year ago I wrote about some of my initial thoughts regarding Joyent's Smart Data Center product and SmartOS in general. A lot has changed since then and the product has both matured and found more acceptance.

I never really got into the "Why" of using the product. We make decisions on product use entirely based on technical and strategic directions. Our use of virtual machine technology is not as common as what one would see among Amazon AWS or Joyent Public Cloud customers. Rather, the requirements have consisted of large, non-transient VMs used for simulations, CAD (of the chip variety) layout, and large data manipulation. We provide hardware for VMs that each require 16-64GBs of RAM, dozens of gigabytes of local storage, and easily 12-16 cores a piece. Up to now, the solutions chosen in academia have been the likes of VMware ESX, Xen, etc. The problem is that the performance, stability, and data integrity requirements either tend to the more expensive end of the product matrix of the above. Perhaps a novel scale-out cloud VM solution on premise would work out better. This second option can be found in either Joyent, which is ready to go now, or OpenStack and friends which will some day achieve similar levels of maturity and flexibility.

Not everything is rosy in Joyent land though. Its primary focus here to now has been to match Amazon AWS in most customer requirements, inclusive of a multitude of transient VMs. This has left the 6.5 revision of the product wanting in areas such as conversion/migration of existing thirty party VMs into the cloud instance, migrating VMs and settings/state between compute nodes, and even migration of head nodes between hardware. Over time, and with help from staff at Joyent, I've worked my way around these edge cases utilizing dataset and package templating, low level use of ZFS snapshots and send/recv, and sometimes just plain old reinstallation of components.

With the announcement of Joyent 7, most of the above issues are being addressed, and we hope to both utilize the newer version and push for more change down the line to make this our go-to tool for virtualization of our entire environment. Where we best hope to help it is in the network space, as we have an obvious preference for the adoption of OpenFlow (SDN) to enable ease of multi-datacenter deployments.

2013 looks like a good year for our VM directions, and I expect others out there will see similar benefit if they just give this technology a try. I didn't give Joyent much thought until the KVM port. A year later, I'm glad we did.

WD Caviar Green drives and ZFS (UPDATED)

We are in the process of outfitting a new primary storage system, and I was of the mind to buy more WD Caviar Green drives, specially more of the 1.5TB WDEADS drives, as we had 4 new ones already that were tested behind a slower RAID card. Before buying more, I searched the usual suspects for pricing, and found the 1TB to 2TB versions of this drive are all priced very well, even for 5400RPM drives, but they now note on different sites and/or comments that they should not be used in RAID configurations. Hmm.

I did a little more research and saw this blog post depicting how one should avoid directly integrating these drives with ZFS. I got a couple, so I decided to put them in my server with an LSI-3442E SAS backplane and tested them. First, I tested my 500GB drives in a mirror set, and doing a "ptime dd if=/dev/zero of=test1G bs=4k count=250000" on the ZFS volume made up of those drives, I transferred 1GB in 3.63 seconds, or 282MB/sec. I then immediately tried the same on my mirror set of the WD drives, benefitting from caching of the first write. After 50+ minutes of waiting, I killed the write and saw that I transferred only 426MB, at a rate of 136KB/sec.

Yes, I can confirm that these drives are less than useless in a ZFS system (see update below), even as a simple two disk mirror set. Some basic iostat showed way too much "asvc_t" service time on the disks, running from 3.5 secs to 10 secs per write, where as the service times for the working 500GB drives were 0.7msec or the like. I had various errors mpt_handle_event_sync errors in my kernel logs, so perhaps there is some specific pathology between the SAS HBA, the SAS/SATA backplane, and these disks. However, we've proven this box works well with various drives. I'm going to try yet another 1.5TB drive, likely the previously maligned Seagate drives, since I've yet to have trouble with the latest firmware on those. My 4 WD drives will be placed in enclosures for external Time Machine backups in the near future. WD Caviar Green != Enterprise RAID drives.

UPDATE:

I'm leaving the above as is, but I think I have discovered perhaps a bad drive in the set, as when I employee 4 drives of this type I saw odd I/O patterns but ok performance in a straight RAID 0. However, I regularly have at least one drive with higher average service times, and trailing I/O writes as it catches up to the other drives. If I have these 4 drives in a pool (RAID 0), I got 193MB/sec writes, and 242MB/sec reads. Sticking them into a RAID10 (2 data, 2 mirror), I got a mirror 78MB/sec writes and 278MB/sec reads.

Splitting them off into two separate RAID1 data pools, I ran my tests and still saw high service times on the drives (only 65 or so, much better than the above, but still slow). Per mirror set performance was dismal, as I regularly got the 150MB/sec+ from a mirror of Caviar Black, but these drives got me just hit 31-34MB/sec (ie, half of the above RAID10). I guess with enough drives I'll get to better numbers in RAID10. In a RAIDZ1 (RAID5) grouping, it was 60MB/sec on the writes, and 172MB/sec on the reads.

So what accounts for the dismal performance I originally saw? I think it has to do with when multiple pools are active, and they are not all of this drive type. My original test had a Hitachi drive set as well as a WD Caviar Green drive set. Although my tests ran one at a time, I'm guessing there was some bad timing/driver issues and/or hardware issues when dealing with the mixed HD media.

A second, update conclusion is that you can use these drives, if only these drive types, in an array. RAID10 will get you sufficient performance, but otherwise you'll want to leave this to secondary storage. Future drive replacement scenarios are a real cause for concern.

ZFS Log Devices: A Review of the DDRdrive X1

My previous notes here have covered the trends to commodity storage, my happiness with most things ZFS and Nexenta, and how someday this will all make for a great primary storage story. At Stanford, we have a lot of disk-to-disk backup storage based on Nexenta solutions, using iSCSI or direct attached storage. We have also had some primary tier uses, but have had to play fast and loose with ZFS to get comparable performance. In essence, we sacrificed some of the ensured data integrity of ZFS to meet end users expectations of what file servers provide.

A typical thing that was done was to set these values:

set zfs:zil_disable = 1

set zfs:zfs_nocacheflush = 1

These flags allowed a ZFS appliance to perform similarly to Linux or other systems when it came to NFS server performance. When you are writing a lot of large files, the ZFS Intent Log's additional latency doesn't affect NFS client performance. However, when these same clients expect their fsyncs to be honored on the back end with mixed file sizes that trend to a large volume of small writes, we start to see pathologically poor performance with the ZIL enabled. We can measure the performance at 400KB/sec in some of my basic synthetic tests. With the ZIL disabled, I generally got 3-5MB/sec or so, or 10x the performance. That's cheating and not so safe if the client thinks a write is complete but the backend server doesn't commit it before power loss or crash.

One ray of hope previously mentioned on this site was the Gigabyte i-RAM. This battery backed SATA-I solution held some promise, but at the time I used it I found a few difficulties. First, the state of the art at that time did not allow removal of log (ZIL-dedicated) devices from pools. One had to recreate a pool if the log device failed. That raised some problems with the i-RAM. First, I had it go offline twice requiring resetting the device, essentially blanking it out and requiring re-initializing it as a drive with ZFS. Second, the connection was SATA-I only, with it not playing well with certain SATA-II chipsets or mixed with SATA-II devices. Many users had to enable it in IDE mode versus the preferred AHCI mode.

Time has passed, and new solutions present themselves. First, log devices can be added or removed from a pool at any time, on the fly. Also new to the discussion is the DDRdrive X1 product. This mixed RAM and NAND device provides for a 4G drive image with extremely high IOPS and a solution to save to stable store (NAND SLC flash) if power is lost on the PCI bus. The device itself is connected to a PCI-Express bus, with drivers for OpenSolaris/Nexenta (among others) that make it visible as a SCSI device.

I tried different scenarios with this ZIL device, and all of them make it a sweet little device. I had mixed files that I pushed onto the appliance via NFS (linux client) and found that I could multiply the number of clients and linearly increase performance. Where I would hit 450KB/sec without the ZIL device but not improve that rate by much with additional writers of data, using the ZIL log device immediately resulted in a good 7MB/sec of performance, with 4 concurrent write jobs yielding 27MB/sec. During this test, my X1 showed only a 20% busy rate using iostat. It would appear that I should get up to 135MB/sec at this rate (5x the concurrent writers), but my network connection was just gig-e, so getting anywhere near 120+MB/sec would be phenomenal. Another sample of mixed files with 5 concurrent writers pushed the non-X1 config to 1.5MB/sec, but in this case, the X1 took my performance numbers to 45-50MB/sec.

So what is providing all this performance? As I mentioned above, the fsyncs on writes from the NFS client enforce synchronous transactions in ZFS when the ZIL is not disabled. My IOPS (I/O Operations per second) without a X1 log device were measured around 120 IOPS. With the dedicated RAM/NAND DDRdrive X1 solution, I easily approach 5000 IOPS. Those commits happen quickly, with the final stable store to your disk array laid out in your more typical 128K blocks per IOP. This dedicated ZIL device has been shown to do up to 200000 IOPS in synthetic benchmarks. Lets try the NFS case one more time, in a somewhat more practical test.

Commonly, in simulation, CAD applications, software development, or the like you will be conversing with the file server committing hundreds to thousands of small file writes. To test this out and make it the worse case scenario of disk block-sized files, I created a directory of 1000 512 byte files on the clients local disk. I did multiple runs to make sure this fit in memory so that we were measuring file server write performance. I then ran 400 concurrent jobs writing this to the file server into separate target directories. First, with the dedicated ZIL device enabled, I got 24MB/sec write rates averaging 6000 IOPS. I did spike up to 43K IOPS and 35MB/sec, likely when committing some of the metadata associated with all these files and directories. Still, the X1 was only averaging 20% busy during this test.

Next, I disabled the DDRdrive X1 and tried again, hitting the same old wall. This was the pathological case. With 400 concurrent writes I still just got 120 IOPS and 450KB/sec. My only thought at the time was "sad, very sad".

You can draw your own conclusions from this mostly not-too-scientific test. For me, I now know of an affordable device that has none of the drawbacks (4K block size, wear leveling) of SSD drives for use as a ZIL device. One can now put together a commodity storage solution with this and Nexenta, and have the same expected performance without compromise as one would expect from any first tier storage platform.

That leads me to the "one more thing" category. I decided to place some ESX NFS storage-pooled volumes on this box, and compare it to the performance of the NetApps we use to manage our ESX VMs (NFS). The file access modes of the VMs tend to be similar to mixed size file operations, but they do tend to be larger writes so the ZIL may not have as drastic of an effect. Anyway, I tried it without the X1 and I got 30-40MB/sec measured disk performance from operations within the VM (random tests, dd, etc). Enabling the ZIL device, I got 90-120MB/sec rates, so we still got a 3x improvement. I couldn't easily isolate all traffic away from my NetApps, but I averaged 65MB/sec on those tests.

Here, I think the conclusion I can draw is this: The dedicated ZIL device again improved performance up to matching what I theoretically can get from my network path. The comparison one can safely make with a NetApp is not that its faster, as my test ran under different loads, but that it likely can match the line rates of your hardware and remove from the equation any concern for filesystem and disk array performance. Perhaps in a 10G network environment or with some link aggregation we can start to stress the DDRdrive X1, but for now its obvious that it enables commodity storage solutions to meet typical NAS performance expectations.

ZFS Resilver quirks and how it lies

One of my ZFS-based storage appliance was running low on disk space, and since I made it a three way stripe of mirrored disks, I could take the 6 500GB drives and replace them with 1.5TB drives each in place, with the result a major increase in capacity. Nifty ZFS software RAID feature versus typical hardware RAID setups. Its all good in theory, but resilvering (rebuilding an array pair) after replacing a drive takes quite some time. Even with only about 400GB to rebuild per drive, one sees the resilvering process cover 90% of the rebuild in 12 hours or so, but that last 10% takes another 10-12 hours. I think this has a lot to do with how snapshots or small files hurt ZFS performance, especially when you are close to a full disk. But its all just as guest as to why its slow on the tail end.

The resilver went as planned, replacing one drive after another serially, but taking care to only do one drive of a pair at a time. Near the end, I started to get greedy. With 98% done on one resilver, I detached a drive in another mirrored pair on the same volume, planning on at least placing the new drive into the chassis so I could start the final drive resilver remotely. To my surprise, the resilver restarted from scratch, so I had another 24 hours of delay to go. So, any ZFS drive removals will reset in progress scrubs/resilvers!

I then decided just to go ahead with the second resilver. This is where it got really strange. The two mirrored pairs started to resilver, and the speed was seemingly faster. After 12 hours, both pairs had about 400GB resilvered and the status of the volume indicated it was 100% done and was finishing. Hours later, it was still at 100%, but the resilver counter per drive kept climbing. Finally, after the more typical 24 hours or so, it noted it was completed.

  pool: data
state: ONLINE
scrub: resilver completed after 26h39m with 0 errors on Tue Nov 24 22:33:46 2009
config:

 NAME        STATE     READ WRITE CKSUM
 data        ONLINE       0     0     0
   mirror    ONLINE       0     0     0
     c2t1d0  ONLINE       0     0     0
     c2t0d0  ONLINE       0     0     0
   mirror    ONLINE       0     0     0
     c2t3d0  ONLINE       0     0     0
     c2t2d0  ONLINE       0     0     0  783G resilvered
   mirror    ONLINE       0     0     0
     c2t5d0  ONLINE       0     0     0
     c2t4d0  ONLINE       0     0     0  781G resilvered

Yes, it looks like at least with this B104+ kernel in NexentaStor, the resilver counters lie. When you have two ongoing resilvers, each counter is nominally the total data resilvered across the whole pool. You'll thus need to wait for double the expected data amount before it completes. Thus, its very important to not reset the system until 100% turns into a "resilver completed..." statement in the status report.

Amanda: simple ZFS backup or S3

When I first started researching ZFS, I found it somewhat troubling that no native backup solution existed. Of course there was the ZFS send/recv commands, but those didn't necessarily work well with existing backup technologies. At the same time, the venerable open source backup solution, amanda had found a way to move beyond its limitation of maximum tape size restricting backup run size. Over time, we have found ways to marry these two solutions.

In my multi-tier use of ZFS for backup, I always need an n-tier component that will allow for permanent archiving to tape every 6 months or year, as deemed fit for the data being backed up. These are full backups only, and due to the large amounts of data in the second tier pool, a backup to tape may span dozens of tapes and run multiple days. I found I had to tweak amanda's typical configuration to allow for very long estimate times, as the correct approach to backing up a ZFS filesystem today involves tar. Amanda's approach does a full tar estimate of a backup before a real backup is attempted. Otherwise, a sufficiently tape library is all you need and a working amanda client configuration on your ZFS-enabled system.

For those following along, I'm an avid user of NexentaStor for my second tier storage solution. Setup of an amanda client on that software appliance is actually quite easy.

setup network service amanda-client edit-settings
setup network service amanda-client conf-check
setup network service amanda-client enable

That's all that one needs to do. There is a sample line in the amanda configuration that you adjust in the first command above. The line I used is similar to this:

amandasrv.stanford.edu amanda amdump

You'll find that depending on your build of amanda server, that you'll either have the legacy user name of "amanda", the zmanda default of "amanda_backup", or the Redhat default of "backup" as the user things run as. I guess there had to be a user naming conflict at some point with "amanda".

The hardest part of the configuration is finding where you have your long term snapshots. Since a backup run can take days to weeks, you'll likely wish to backup volumes relative to a monthly snapshot. In your amanda /etc/amanda/CONFIDR/disklist configuration, a sample you may have for a ZFS-based client named nexenta-nas with volume tier2/dir* is:

nexenta-nas /volumes/tier2/dir1/.zfs/snapshot/snap-monthly-1-latest user-tar-span
nexenta-nas /volumes/tier2/dir2/.zfs/snapshot/snap-monthly-1-latest user-tar-span


Note well the use of user-tar-span in the two lines above. This allows for the backing up large volumes over multiple tapes in amanda. That one limitation of tape spanning in amanda was solved in a novel way. They break up backup streams into "chunksizes" of a set size to allow for a write failure at the end of one tape to begin fresh again at the beginning of that chunk on the following tape. This feature allows amanda to also be used to backup to Amazon's S3 service. Yes, instead of going to tape, you can configure a tape server to write to an S3 service. S3 limits writes to a maximum of 2GB a file, and amanda's virtual tape solution combined with that chunk sizing of backups works wonderfully to mate ZFS-based storage solutions to S3 for an n-tier solution. Please consult Zmanda's howto for configuring your server correctly. There really is nothing left to configure to get ZFS data to S3.

Recommended Disk Controllers for ZFS

Since I've been using OpenSolaris and ZFS (via NexentaStor, plug plug) extensively, I get a lot of emails asking about what hardware works best. There have been various postings on the opensolaris and zfs lists to the same effect. A lot of people reference the OpenSolaris HCL lists which leave the average user scratching their head with more questions than answers. More to the point, the HCL doesn't tend to answer the more direct question of what hardware should I get to build a ZFS box, NAS, etc. Its important to note that in the case of ZFS, all that extra checksum, fault management, and performance goodness can be negated by selecting a "supported" hardware RAID card. Worse yet, many RAID cards are not fully interchangeable on the spot. What do you want for ZFS?

First, pick any 64-bit dual core or better motherboard or processor. If you can get ICH6+, nvidia, or Si3124-based on board SATA, then you are in good shape for your basic ZFS box with on-board SATA for your system disks alone. System disk can tend to be low 5400RPM 2.5 inch SATA-I drives. Many people then desire some large memory, battery-backed RAID card, and my tests with the high end LSI SAS cards show that memory on the RAID card doesn't do you as much good as having a recipe of lots of system RAM, a sufficient number of cores, many disk drives for the spindles, and sufficient use of the PCIX/PCIe bus using JBOD only disk controllers. I'll cover the controllers next, but I'd recommend at this point 4GB of RAM minimum, dual core at greater than 2ghz, and for any good load, at least two PCI-X or multi-lane PCIe card.

Disk controllers are where the real questions are asked. Over multiples iterations, heavy use, and some anecdotal evidence, we are down to some sweet spots. For PCI-X, there is one game in town, the Marvell-based AOC-SATA2-MV8, used in the X4500. At $100 for 8 JBOD SATA-II ports, it just works and is fault managed. Stick just SATA-II disks on these, and keep any SATA-I disks on the motherboard SATA ports for system disks. I'll add that various Si3124 based cards exist here, but not for sufficient port density.

SuperMicro AOC-SATA2-MV8 link

When it comes to PCIe, there isn't any good high port count options for SATA. If you need just 2 ports, or eSATA, there are various solutions based on the Si3124 chipset, and SIIG makes many of them for $50 each. However, in the PCIe world, the real answer is SAS HBAs that connect to internal or external mixed SAS/SATA disk chassis. Again, most SAS HBAs are either full fledged RAID without JBOD support, or simply don't work in the OpenSolaris ecosystem. 3ware is a lot cause here. The true winner for both cost and performance, while providing the JBOD you want, is the LSI SAS3442E-R.

CDW catalog link for LSI 3442ER
LSI 3442ER product page

Its $250, but I've seen it as low as $130. 8 channels, with both 2 internal ports (generally 8 drives are connected to a single SAS port) as well as the external port. You can use this with an external SAS-backed array of SATA drives from Promise, for instance, to easily populate 16 or 32 drives internally, with an additional 48 drives externally, just from the one card. Would I suggest that many on that single card? No, but you can. Loading up your system with 2 or 4 of these cards, which are based on the LSI 1068 chipset that is well supported by Sun is the best way forward for scale out performance. I was given some numbers of 200MB/sec writes and 400MB/sec reads on an example 12-drive system using RAIDZ. Good numbers, as I got 600MB/sec reads on a 48-drive X4500 thumper.

If you have PCI-X, go Marvell. PCIe? Go LSI, but stick to the JBOD-capable not-so-RAID HBAs. Don't just trust me, throw a $100 or two at these and try it yourself. You'll see a better investment than $800 at the larger RAID cards. I went the latter route and have paid dearly (Adaptec, LSI, you name it). What worked from the beginning and is working today are the Marvell cards here, and I've been playing with new systems that use the LSI 3442ER.

Mixing SATA dos and donts

Another day, another bug seemingly hit. I've known for some time that mixing SATA-I and SATA-II devices on the same controller with regards to OpenSolaris seems to be unwise. I've already had systems with the initial ZFS-boot drive being a small capacity and thus likely SATA-I, but the data volumes were SATA-II. My recent issues with the iRAM could be related to having a SATA-I device after a SATA-II drive in the chain, but nothing has been concrete.

However, today I discovered something else. One array I have is made up of all SATA-I drives and was used by a SATA-I RAID card that went south. I happily replaced it with the Marvell SATA-II JBOD card, and it was working just fine. I then lost the 6th of 7 drives, and went back to the manufacturer to try and buy a replacement. Sadly, these Raid Edition drives have been "updated" to be at a minimum of SATA-II for the same model. Replacing the failed SATA-I with the SATA-II worked, but on subsequent reboots, the 7th drive tended to not be enumerated by the Marvell card at startup, and even after re-inserting it, a "cfgadm" was necessary to activate it. Even then, a "zpool import" or "format" to introspect the now configured drive would wedge and never complete the command. Weird, right?

The solution to return to stability was to swap the 6th and 7th drive, so that the SATA-II disk came after all the SATA-I devices in the chain. I'm not sure why it works, but every reboot works now, it never fails to enumerate that last drive and there is no need to manually cfgadm configure the drive post boot. Therefore, a set of truisms are starting to come together with mixed SATA drives. Whether Marvell, Sil3124, or the like, its never a good idea to mix SATA-I and SATA-II devices on a single controller, but if necessary, make sure that the SATA-II drives come after the SATA-I drives. The best configuration is to restrict SATA-I boot devices, such as small 5400 "laptop" drives to their own onboard SATA interface, and leave all SATA-II devices to add-on boards.

The problem with slogs (How I lost everything!)...

A while back, I spoke of the virtues of using a slog device with ZFS. The system I went into production with had an Nvidia-based SATA controller onboard and a Gigabyte i-RAM card. No problems there, but at the time it was a cmdk driver (PATA mode) for my OpenSolaris-based NexentaStor NAS. After a while, I got an error where the i-RAM "reset" and the log went degraded. The system simple started to use the data disks for the intent log. So, no harm done. Its important to note that the kernel was a B70 OpenSolaris build.

Later, I wanted to upgrade to NexentaStor 1.0, which had B85. Post upgrade or using a boot CD, it would never come up with the i-RAM attached. The newer kernel was an nv_sata driver, and I could always get it to work in B70, so I reverted to that. This is one nice feature that Nexenta has had for quite some time, in that the whole OS is checkpointed using ZFS to allow reversions if an upgrade doesn't take. Well, the NAS doesn't like having a degraded volume, so I've been trying to "fix" the log device. Currently, in ZFS, log devices cannot be removed, but only replaced. So, I tried to replace it using the "zpool" command. Replacing the failed log with itself always fails as its "currently in use by another zpool". I figured out a way around that, and that was to fully clear the log using something like "dd if=/dev/zero of=/dev/mylogdevice bs=64k". I was able to upgrade my system to B85, and then I attempted to replace the drive again, and it looked like it was working:

  pool: data
 state: DEGRADED
status: One or more devices is currently being resilvered.  The pool will
 continue to function, possibly in a degraded state.
action: Wait for the resilver to complete.
 scrub: resilver in progress for 0h0m, 0.00% done, 450151253h54m to go
config:

 NAME         STATE     READ WRITE CKSUM
 data         DEGRADED     0     0     0
   raidz1     ONLINE       0     0     0
     c7t0d0   ONLINE       0     0     0
     c7t1d0   ONLINE       0     0     0
     c6t1d0   ONLINE       0     0     0
 logs         DEGRADED     0     0     0
   replacing  DEGRADED     0     0     0
     c5d0     UNAVAIL      0     0     0  cannot open
     c8t1d0   ONLINE       0     0     0

errors: No known data errors

Note well, that it is replacing one log device with another (using the new nv_sata naming). However, after it reached 1% it would always restart the resilver with no ZFS activity, no snapshots, etc. The system was busy resilvering and resetting, getting no where. I decided to reboot to B70, and as soon as that came up, it started to resilver immediately and it proceeded after quite a long time for a 2GB drive to complete the resilver. So, everything was now fine, right?

This is where things really went wrong. At the end of the resilver, it still considered the volume degraded, and looked like the above output but with only one log device. Rebooting the system, the volume started spewing out ZFS errors, and checksums counters went flying. My pool went offline. Another reboot, this time with the log device disconnected due to nv_sata not wanting it connected for booting purposes causes immediate kernel panics. What the hell was going on? Using the boot cd, I tried to import the volume. It told me that the volume had insufficient devices. A log device shouldn't be necessary for operation, as it hadn't needed it before. I attached the log device and ran cfgadm to configure it, which works and gets around the boot time nv_sata/i-RAM issue. Now it told me that I have sufficient devices, but what happened next was worse. The output showed that my volume consisted of one RAIDZ, an empty log device definition, and additionally my i-RAM as an additional degraded drive added to the array as a stripe! No ZFS command was run here. It was simply the state of the system relative to what the previous resilver had accomplished.

Any attempt to import the volume fails with a ZFS error regarding its inability to "iterate all the filesystems" or something to that affect. I was able to mount various ZFS volumes read-only by using the "zfs mount -o ro data/proj" or similar. I then brought up my network and manually had to transfer the files off to recover, but this pool is now dead to the world.

What lessons have I learned? Slog devices in ZFS, though a great feature, should not be used in production until they can be evacuated. There may be errors in the actions I took above, but bugs that I see include the inability for the nv_sata driver to deal with the i-RAM device for some odd reason, at least in B82 and B85 (as I've so far tested). The other bug is that a log replace appears to either not resilver at all (B85) or, when resilvering in older releases, causes the system to not correctly resilver the log but instead to shim the slog in as a data stripe. I simply can't see how that is by any stretch of the imagination by design.

ZFS: Is the ZIL always safe?

One of my ZFS-based appliances, used for long term backup, was upgraded from B70 to B85 of OpenSolaris two weeks ago. This time around, I re-installed the system to get RAIDZ2, and certain "hacks" that I've been using were no longer in place. The old settings were in /etc/system, and are the well known zil_disable and zfs_nocacheflush enabling. They were left there from when the system temporarily acted as a primary server for a short time with its Adaptec (aac) SATA RAID card and its accompanying SATA-I drives. Since the unit was UPS attached, it was relatively safe for NFS client access, and later on there was no direct client access over NFS. No harm done, and stable for quite some time over multiple upgrades from B36 or so, over a year without an error.

A curious thing happened as soon as I upgraded without these somewhat unsafe settings for the kernel. I started to get tons of errors and twice my pool as gone completely offline until I cleared and scrubbed it. An example of the errors:

NAME STATE READ WRITE CKSUM
tier2 DEGRADED 0 0 0
raidz2 DEGRADED 0 0 0
c1t1d0 FAULTED 0 64 0 too many errors
c1t2d0 DEGRADED 0 46 0 too many errors
c1t3d0 DEGRADED 0 32 0 too many errors
c1t4d0 DEGRADED 0 47 0 too many errors
c1t5d0 DEGRADED 0 39 0 too many errors
c1t6d0 FAULTED 0 118 0 too many errors
c1t7d0 DEGRADED 0 57 0 too many errors


Nothing explained the turnaround from stable to useless for any writes. I also got some read errors, and no nightly rsync against this tree would survive without incrementing some error count. Was it somehow one of my cache settings on the adaptec card that conflicted with a new version of the "aac" driver? I thought I would need to isolate it, revert perhaps, or consider that somehow my card was simply dying. Perhaps the cache/RAM on the card itself was toast.

A recent post on the opensolaris-discuss mailing lists gave me an idea. Mike DeMarco suggested to a user suffering from repeated crashes that corrupt ZFS until cleared to try and use zil_disable to test "if zfs write cache of many small files on large FS is causing the problems." Makes some sense if the card is somehow trashing on small writes. The use of it for backup means that its being read and written to via rsync and can involve many small updates. I also had various read errors pop up. So, I put the old faithful zil_disable and for good measure the zfs_nocacheflush back after another degraded pool, and after a reboot and scrub, let it do its nightly multi-terabyte delta rsyncs. After a few days, there are no errors. Have I stumbled onto some code path bug that is ameliorated by these kernel options? Do newer kernels have suspect aac drivers?

Perhaps someone will prove the logic of the above all wrong, but for now, I'm returning to the old standby "unsafe" kernel options to keep my pool stable.

Using the iRam: Improving ZFS perceived transaction latency

I've been long overdue in reviewing the Gigabyte iRam card and its affect on performance of your favorite ZFS NAS product. NexentaStor already supports log devices, so the time appeared right to get one of these for a client I consult with to help deal with the noticeable pauses one can see when heavy reads and writes compete on a ZFS pool. I hope that the single threaded nature of those commits is resolved at some future point, but the iRam card appears to be a simple way to inject an NVRAM-like device into your commodity NAS solution.

The card itself is simply four DIMM sockets for DDR RAM, with a battery backup, reset switch, power driven from a PCI bus, and a single SATA-I connection to plug the unit into your existing SATA interfaces. Already you can see that the performance limit is 150MB/sec based on the SATA-I spec. What does this card do though? Near instant reads and writes in a safe battery-backed ramdisk that your system sees as a 2GB or 4GB drive, just what you'd want for a dedicated write commit device. In the case of many spindles in an array, you likely can do better than this device for true performance, but in the case of many small commits, the near perfect latency of RAM is much more ideal to keep writes happening without stalling the drives for reads. Since its a "slog" device by ZFS terms, it will regularly commit to the real underlying storage at full disk bandwidth. Therefore, even when writes must compete with reads on the physical disk, you limit your exposure to perceived stalls in I/O request even in the higher load cases.

For my non-production test, I actually put together the worse case scenario: An iSCSI backed ZFS array with NFS clients and many small files. In this case, any NFS writes require 3 fsyncs on the back end storage as required by NFS (create,modify,close). This is actually similar to CAD libraries, which the test was made to reflect. Using iSCSI devices, you can inflict much higher latencies. My iSCSI targets are actually older SATA-I drives themselves on a SBEi Linux based target using 3ware 8500s. Again, no where near ideal.

Creating a directory of 5000 small 8k files, I copied this from a linux gig-e connected client to a ZFS pool (made of two non-striped iSCSI luns), and got a meager 200K/sec write performance over NFS. If I stripe the data instead in the ZFS pool, I increased the numbers to 600K/sec at some points. Adding a 2GB Gigabyte iRam drive, I increased those numbers up to 9MB/sec, but averaging around 5MB/sec overall. That's at least 10 times the performance. Again, this test involves many i/o operations instead of using any bandwidth.

How fast can data be written to and read from that log device? My tests showed that 100MB/sec for reads and writes were common, with writes only bursting to those numbers for larger streaming data sets. In the case of the iSCSI nodes in question, each one could be pulled at a top rate of 45MB/sec, but averaging closer to 27MB/sec. Nominally, you can see that we are 3x better than at least these gig-e iSCSI devices.

The final production installation of the iRam device was with a SATA-II DAS array, and even in heavier load scenarios, we saw the wait cycle for write commits to the drives limited, and a steady 100+MB/sec use of the commit log (reads and writes). The only caveat for using such a device is that the current builds of OpenSolaris and thus NexentaStor do not allow you to remove it once added to a pool. A future release is supposed to address that.

Swept Under the Rug

In our day to day management of technology, we tend to pick paths that resolve the most pressing pain points. Inadvertently, we often also sweep certain problems under the rug, awaiting the day when it all must be cleaned up. Many choices do exactly this, solving the present problem while creating perhaps larger problems down the road. In my evolving strategy on storage, the move away from tape to disk-based online storage solves the most obvious problems but creates a whole series of other problems, including file based disaster recovery, long term maintenance of the underlying disk technology, true long term persistence of data, and general accessibility of the data by future technology. Today, I'll focus only on our next major pain point, disaster recovery.

Recently, a few instances occurred that underscored the need for better thought out solutions than what we already put in place. We thought we may be ahead of the curve with tiered copies of data on secondary NAS solutions, with our backup windows well within reason. Its obvious we made the right choice in doing incremental file based backups to secondary NAS, as the end data containers are universal across network file protocols. Recovery of any given file or perhaps even full data store recovery still beats that of tape libraries multiple times over. However, the architecture in place has allowed us to scale from the gigabyte world to the terabyte world. Our backup windows are well in hand, and spot recovery is a cinch. But there are some problematic disaster recovery scenarios.

The first scenario was just felt a week ago. A mere 50GB file store of Maildir formatted mail, where each message is a file itself, with mail folders represented by directories, had write errors on its underlying Linux XFS volume. This is by far not our largest install of such. Various mail servers for separate organizations we deal with are over 500GB in size. We suspect the RAID card's NVRAM was toast, disallowing further writes, and we had to migrate the mail to another server quickly. Simple enough, let's recover from our second tier mail store, right? The attempt was made, but we found ourselves limited not by the reading of millions of small 1K files so much as recommitting those files onto a journaled filesystem. The metadata updates of the files alone were bad enough. In the end, we were limited by file operations per second, and not pure bandwidth to the disk. Our estimated time of recovery was a minimum of 14 hours, and only for 50GB. A clue to the long term solution to this was in how we restored everything in less than 2 hours. In this case, we relied upon an xfsdump from the read-only failing array to a new filesystem on the spare hardware.

The obvious up front answer to disaster recovery of data in a multi-terabyte world is to make sure you have copies of everything in as close to a high availability setup as you can afford. If the underlying RAID array was actually two arrays, with software mirroring across the two, or if it were two separate machines that either attached to shared mirrored arrays or otherwise mirror their underlying RAID arrays over the network, we'd all just worry about natural disasters. Preventing the true disaster recovery scenario up front is the only true way to win, but most of us simply don't have the luxury, the resources, or the ability to safely migrate the myriad of production or otherwise in use solutions over to the ideal configuration. We can all try to reach this nirvana, but its simply not as attainable to most of us as we'd like.

We can, however, address some of the pain of the disaster recovery scenario from disk based solutions. The iSCSI and SAN vendors have been on this for some time, and have extolled the virtues of block based storage. Using such, you can stream I/O at near the theoretical limits of the hardware. However, running all your systems against a SAN throws you down the path of the usual hardware based solution to a general problem, with the usual vendor lock-in quibbles. We already have bought into the software based approach that NexentaStor has offered us, and happily, they already provide a similar solution to our needs. With thin-provisioning of virtualized storage volumes (zvols), one can expose block level storage to clients but still treat them as snapshot capable files, use file level services and such on the back end second level NAS. The clients will generally access these through iSCSI, and they can either directly depend on these network-based volumes as if they were local filesystems, or simply use their filesystem native dump programs to periodically maintain a near synchronized copy of a true DAS filesystem to a second tier block level copy. The latter is nice as it doesn't place undue strain on the back end storage architecture to service all clients in parallel at the fullest of performance for production. We just use network and storage resources for backup.

What does this solve? In the case of the disaster recovery, the reverse backup process can be done, getting streaming I/O rates, perhaps as high as 100+MB/sec over gigabit ethernet when the local arrays fail. In the case of my mail spool filesystem, we recovered at rates of 25-30MB/sec instead of the 500-800K/sec we saw. Even if its not the most up to date copy, if one also did file-level backup of the underlying file system to NexentaStor or the like at a faster interval, you can recover from that incrementally after the first block level recovery. Either way, you taste the sweetness of success. Again, the dirt is stuffed under the rug if this is hundreds of terabytes, and some day soon that may also be just as common place, but perhaps we are again ahead of the curve. The one side affect is that you'll want more cheap storage readily available on the second tier.

I'll quickly describe the second scenario, where a failing system also needed its 1TB of mainly larger files migrated. We saw that our top rates of file level recovery at best were 1GB/minute, but generally less. Again, it would have made sense to have been redundant up front, but the same solution above could more than double the rate of recovery if we could restore the primary file system at block speeds. This is similar to how virtual machines are managed from SAN, iSCSI, or even NFS. The VMs themselves are represented as files, and so operations on these files approach the maximum speed of block storage operations. However, having them on that NAS allows ease of sharing and management, including snapshots. No hardware tricks, all software. We haven't addressed the next stumbling blocks, which include kernel page size limitations on true file I/O, but the dirt is nicely hidden for the time being.

Old is new again

Prior to my happiness in using all things ZFS, I was eagerly testing Lustre. Primary concerns there was both management of the backend storage and ability to translate this network file system to standard NFS and CIFS for non-Linux clients. Cluster FileSystems was recently announced to be acquired by Sun, but prior to that, they made announcements about how Lustre and ZFS would be married. I always thought that perhaps it was that ZFS would be the exposed layer on a client, and that it was the same old OSTs and OSDs on the backend. Not so.

CFS Moves Lustre to ZFS

Instead, ZFS is the backend to the storage nodes. Albeit that the driver pool is somewhat less for OpenSolaris than Linux, it does make one wonder exactly what new flexibility is afforded by this arrangement. Also, how much of this will end up as open sourced code to be incorporated into highly manageable product, such as my new favorite appliance, NexentaStor? I'll be diving further into this as more details emerge. Perhaps others can chime in with more info or clarification.