"OS6" - a question or three

garyd9 wrote:

If I took that older Pro BE box (not Pro 6), swapped out the CPU with an E6700 (which seems to be the highest performing dual core processor known to work on that box regardless of which MB revision is in there), swapped the 1x1GB DDR2 out with 2x2GB sticks, and put OS6 on it (which would void the warranty that's already expired)... would that give me something comparable to the newer 516 box? I realize I'd still be missing some hardware interfaces such as USB3 and eSATA. However, in day to day use as a network attached storage device (and not for things like transcoding) would it be nearly the same as a stock 516? Perhaps a clearer question would be if such a converted box would be representative of what I'd get in purchasing a 516.

You may find the Pro BE as is is good enough for what you want already (what speeds do you get with that?), but if you want to do upgrades you can. So it's up to you if you want to try OS6 if you try it without any upgrades first or do the upgrades then update to OS6. Before replacing the CPU you should make sure you are running the latest BIOS (you can check by looking in bios_ver.log). I would suggest the BIOS upgrade add-on should be installed before upgrading to OS6 as the BIOS upgrade add-on is written for OS4 and won't run on OS6.

Note it is possible to downgrade a legacy ReadyNAS running OS6 back to OS4 but this also requires a factory default.

Apart from the CPU and RAM, the key advantages of the 516 would be the 5 year warranty, the USB3 and eSATA and the HDMI port (for possible future use).

garyd9 wrote:
Jumping back to the dual redundancy and OS6... The method used to get that working with xraid2 is interesting. So set things up with flex-raid raid6 and then just convert it intact to xraid2?

Yes

garyd9 wrote:
Oh, and on the 516, can a eSATA drive be used for backing up the array? Or can they only be used for expanding the array?

I believe you can however do note that some external enclosures may be incompatible.

garyd9 wrote:
For the raid scrubbing on a schedule issue: Can the raid scrubbing be initiated via a command line? If so, would it be feasible to just create a cron job to kick it on a schedule? That would probably suffice as a work-around for a validation check I've become accustomed to.

I think that should be possible. Of course you would want your cron job to check if it's appropriate to run it or not. Blindly running commands like this on a schedule could lead to problems.

garyd9 wrote:
@Mdgm, from the descriptions above and what I've seen in my PM box, it appears that I'd be able to use a newly purchased 516 box to at least get the data off a Radiator4 array in a pinch. Not as clean as a replacement pre-OS6 box (which would just take over the array natively), but at least doable. That's helpful.

Yes

garyd9 wrote:
@StephenB, I appreciate the suggestion to just wait and see what might happen with a new OS6 version, but I'm being pressured to come up with a plan to cycle out the BE's before the end of the year. (It doesn't help that my primary job is software development, so I'm really only able to work on this NAS stuff on my supposedly free time.)

Well if you do get new ReadyNAS units then you will benefit from firmware updates for those boxes such as 6.2. If your Pro BEs still work fine I would recommend keeping them for backup duties at least. With features such as disk spindown (on OS4 and coming in 6.2 for OS6) and the power off/on schedule they needn't draw too much power if you don't want them to.

With the BE's, I'm already saturating gigabit using 5 drives configured for RAID6 via samba. When using AFP, it only gets to about 70%. (That might be due to Mac machine's accessing it, though.) There aren't enough individual users creating heavy usage at any given moment for user counts to be an issue.

The reason for changing the h/w on the BE box before OS6 would only be to bring the differences between the retrofit and an actual 516 to a minimum. I don't want any avoidable shortcomings to be present when testing, as then I wouldn't be convinced that it's a valid test. The lack of USB3 and eSATA wouldn't be missed without actually using those interfaces, so that's a non-issue. I want to get as close as possible to an actual 516 to evaluate the overall package...

I just ordered a refurb'd E6700 (SL9ZF stepping) for $31 (US) shipped, and will find some old DDR2 memory laying around the office. (I should be able to find a couple of 2GB/667 or /800 sticks.) I'll re-use the drives that are already in that box (seagate constellation es.2 1TB drives.) I have the BIOS already updated to v2.0 (with a 2010 year) already. I also have a VGA header somewhere I can use to troubleshoot if the h/w changes cause problems.

As for keeping the ProBE boxes for backup, I agree it would be an intelligent thing to do, but likely won't happen. Don't worry, though -- I'm sure they'll be re-purposed as a backup device for some overworked engineer's personal use. ;)

garyd9 wrote:
With the BE's, I'm already saturating gigabit using 5 drives configured for RAID6 via samba. When using AFP, it only gets to about 70%. (That might be due to Mac machine's accessing it, though.) There aren't enough individual users creating heavy usage at any given moment for user counts to be an issue.

Well what your client hardware is like is also a factor. Do your client machines use SSDs? Do you they use the same kinds of files? A large sequential file will transfer much more quickly than lots of small files summing up to the same total size.

garyd9 wrote:

The reason for changing the h/w on the BE box before OS6 would only be to bring the differences between the retrofit and an actual 516 to a minimum. I don't want any avoidable shortcomings to be present when testing, as then I wouldn't be convinced that it's a valid test. The lack of USB3 and eSATA wouldn't be missed without actually using those interfaces, so that's a non-issue. I want to get as close as possible to an actual 516 to evaluate the overall package...

Then give the upgrades a try.

garyd9 wrote:

I just ordered a refurb'd E6700 (SL9ZF stepping) for $31 (US) shipped, and will find some old DDR2 memory laying around the office. (I should be able to find a couple of 2GB/667 or /800 sticks.) I'll re-use the drives that are already in that box (seagate constellation es.2 1TB drives.) I have the BIOS already updated to v2.0 (with a 2010 year) already. I also have a VGA header somewhere I can use to troubleshoot if the h/w changes cause problems.

Sounds like a good plan

garyd9 wrote:
Don't worry, though -- I'm sure they'll be re-purposed as a backup device for some overworked engineer's personal use. ;)

:)

mdgm wrote:
What I did: 1. Do a factory default 2. Disable X-RAID 3. Delete the volume 4. Create a new RAID-6 volume with the same name 'data' 5. Re-enable X-RAID 6. Create the default shares manually

So, I've started to play around with this. I don't have the replacement processor yet, but I did up the RAM to 4GB. What I did was to get OS6 going with 1x1TB and 3x2TB drives, disabled XRAID, created a RAID6 volume and re-enabled XRAID. (It's currently building the volume.) The reason I did the drives that way is to validate if I can add another 2TB drive and another 1TB drive while maintaining RAID6 and XRAID. It might take a few years for the volume to get out of degraded status. For some reason, I decided to stress the thing out by copying a few hundred GB's of data to a newly created share while it's still rebuilding the array. (On the other hand, I have to admit that I'm impressed that I'm getting 75MB/sec copying to the degraded array while it's rebuilding!)

Some very initial comments... Obviously, these are based on coming from v4 of the firmware:

I really don't like that I have to activate DLNA on an entire share. I much preferred the option of specifying a share AND directory. It's possible that I'd have some "pictures" (for example) that I don't want scanned and available by DLNA, and it seems annoying to have to create multiple shares to accomplish that. As well, I might want to only have a single share that houses all my media types, but specify certain directories within that single share for certain DLNA types. (ie: //DLNA Scanned/Music, //DLNA Scanned/Video, etc)

Please do something different in FV concerning getting drive device data (SMART data.) Those little javascript popup windows are pretty, but they aren't very useful. On my desktop (using Chrome), I have a hard time scrolling the little windows in order to see all the data (I have to move my mouse just perfectly to get inside the window and then use my scroll wheel), and they don't work properly AT ALL on my android devices.

For some reason, "Apps->Available Apps" is completely empty for me (and refresh does nothing.) I'm not sure if that's because the hardware is unsupported, or if it relates to deleting and re-creating the volume when getting RAID6 + XRAID active. (Actually, it can't be the hardware, as there was stuff listed there the FIRST time I booted OS6 after a factory default. I've done a couple of factory defaults since then, but I don't remember when they list became empty...)

I like the "performance" graphs idea. In particular, I could see the "network" graph (and it's various options) being extremely useful.

The snapshot "options" are very different from what I expected. They make more sense after reading the manual (and searching this forum), but I suspect I'm going to be coming back with several questions related to that.


More to come...
Gary

garyd9 wrote:
It might take a few years for the volume to get out of degraded status. For some reason, I decided to stress the thing out by copying a few hundred GB's of data to a newly created share while it's still rebuilding the array. (On the other hand, I have to admit that I'm impressed that I'm getting 75MB/sec copying to the degraded array while it's rebuilding!)

Please note that the resync is a low priority process. However your volume is redundant throughout this time in the sense that if a few disks fail your data remains intact. Obviously copying data while the array is rebuilding will slow the process a bit.

garyd9 wrote:

For some reason, "Apps->Available Apps" is completely empty for me (and refresh does nothing.) I'm not sure if that's because the hardware is unsupported, or if it relates to deleting and re-creating the volume when getting RAID6 + XRAID active. (Actually, it can't be the hardware, as there was stuff listed there the FIRST time I booted OS6 after a factory default. I've done a couple of factory defaults since then, but I don't remember when they list became empty...)

That could be due to incorrect network settings. Do you have Genje Apps enabled? If so try disabling that.

mdgm wrote:
That could be due to incorrect network settings. Do you have Genje Apps enabled? If so try disabling that.

No, genie is disabled. (I did try enabling it at one point to see if it would help, but no luck, so disabled it again.) I _can_ upload .dep files, but dependencies aren't being installed automatically. (Example: I downloaded the transmission package from "http://apps.readynas.com/pages/?page_id=9", and uploaded it.. .but it didn't work. apt-get showed a missing dep of "transmission" for "transmissionr6"

I dropped to SSH and noticed that "apt-get update" is giving a PGP (er.. GPG error: "W: GPG error: http://apt.readynas.com 6.1.9 Release: The following signatures were invalid: BADSIG B7F02108D553259A RAIDiator Release Key <admin@readynas.com>") No idea if that's related or not.

Ah.

That repository might have been updated recently and the changes can take a little while to propagate through.

If that isn't fixed by Monday (i.e. If an apt-get update doesn't work then) let me know.

Which 6.1.9 RC are you running?

mdgm wrote:
Which 6.1.9 RC are you running?

I don't think I'm running an RC. I took the "easy 3 step" method of R4toR6_latest.bin. (Oddly, I can't seem to find a version number in the frontview interface)

Due to important fixes in 6.1.9 RCs such as one relevant to doing a factory default (which is required as part of the R4toR6 upgrade process) I chose to update that to update to upgrade to 6.1.9 RCs.

When you SSH in you should see which RC you are running.

It would also show in the logs in os_version.log

Heh.. I see a slot where there should probably be a firmware version.. but it's blank. On the "system"->"Overview" page. It shows my model, Name, Status, a blank for the AV status, the proper serial number, a blank for the firmware, and the proper device time.

SSH shows RC11