My "activity" light NEVER stops unless I turn off the power. NEVER.
I usually run MYSQL server (for the XBMC database) and the NAS holds my video files. The XBMC accesses the NAS via NFS.
Response from MySQL queries and file requests has been EXTREMELY slow lately. So, to test, I turned off MySQL and NFS, so that nothing can be quietly accessing the NAS.
The light remains nearly solid. I can't run IOTOP, since the kernel provided isn't compiled for that, so I can't determine where the IO is occurring.
Any ideas? Nothing I do seems to slow the intense activity and slowness. There's got to be an answer. I've been pretty dissatisfied with this ReadyNAS unit since I bought it. It's never really lived up to what I thought a NAS should be. I guess that's what I get for buying into the low end of the market.
I would begin to disable all the add ons, the antivirus, the backup jobs and the snapshots. If things improve, I would re-enable them one by one to get a culprit. If things don't improve, I would to a defrag. You could do a write speed test with a 2gb file to see if the NAS under performs (should be around 90-95Mbytes/s in read and 40-50Mbytes/s in write for the 104 over gigabit ethernet)
Come back here with the results of that. Depending on those we might have some other tests to do.
My 104 writes something every 3 to 5 seconds but that's not what you seem to describe.
OK. I sent the logs to the email in the link mdgm provided above.
To the other poster, I've turned off all services. I don't run Antivirus on the NAS since 90% of my content is video and my primary machines are Macs anyway...
I tried running a defrag -- no status update in the web panel. It just acted like it did nothing. The log says "started defrag" and nothing more. I started a scrub, just in case there was some RAID corruption causing a slowdown. As of this post, it's 52% complete (4 4tb drives take a while!).
Reading and writing to the array seem fast enough -- once I can get the file to start. It's like the NAS is too busy to initiate the thread, but once it does, transfers occur normally.
When I SSH into the box and run TOP, it all looks "normal." I'd like to run IOTOP to see where the IO processes are, but apparently that support is not compiled into the kernel (Why not? I would think that being able to monitor IO processes on such low-powered hardware would be a no-brainer!)
Hopefully, the logs will reveal what is causing the crazy amount of IO. I expect the scrub to take until tomorrow sometime to complete, and I will report back then with any changes (or lack thereof) from that process. Hopefully we can get a defrag to work, too...but from my understanding of BTRFS and the fact that I don't do a lot of write IO, I should not have a significant fragmentation issue -- but I'm willing to try it.
The first defrag can take some time (could be up to 3-4 days depending on the fragmentation), afterwards it should complete in a matter of hours (mine takes less than 2h for 5Tb). There is indeed no progress bar, you will just receive a mail at the end.
I agree with you that iotop would have been quite an help, defrag should have a progress bar too as well as backups (or at least the data processed + the amount remaining + the current file). Maybe in an incoming update.
Well, mixed results: The scrub seemed to help with the slow performance issue, but my activity light is still on solid with few breaks. It flickers every now and again, but mostly it is on solid. That's a lot of activity. I await the response from you once you look at the logs to see if there is anything going on..maybe some log file in /var somewhere?
What does it do? The ReadyNAS doesn't seem to be much more than an ARM-based processor with a SATA controller doing software RAID. What is so "special" about readynasd?
That would be contrary to the standard behavior of a disk activity light as has been established by over 30 years of precedent. No activity was always OFF and when there was activity it would blink. So a solid light meant that there was so much activity that you could not see the times when the light was off.
Can we get someone from Netgear to confirm this? Is the activity light supposed to be on SOLID when there is NO activity? If so, please fix that in a firmware update!
on => disk present off => no disk blink => activity
I know nobody reads the manuals anymore...
PS : I should have asked from the beginning if "on" was solid or blinking, I assumed blinking and I assumed that you were hearing disks reading/writing. My bad.
Personally I think that "on" makes sense for "disk present".
If "off" means disk present but not used, and also means "no disk" or "disk failed", then I can't tell if something is wrong. Its much better as its documented.
I'm not talking about the lights labelled 1,2,3,4 that indicate a disk is inserted into a given slot in the RN104. I'm talking about the round blue light that is labelled "act." My lights labelled 1,2,3,4 are always on solid as soon as the RAID array comes online, indicating that the disks in those slots are online and ready. They NEVER blink.
I'm not talking about the lights labelled 1,2,3,4 that indicate a disk is inserted into a given slot in the RN104. I'm talking about the round blue light that is labelled "act." My lights labelled 1,2,3,4 are always on solid as soon as the RAID array comes online, indicating that the disks in those slots are online and ready. They NEVER blink.
yes, you do have a point on "act" - that the individual disk lights ought to be enough. Though xeltros is quoting from the right part of the documentation, and therefore it is "working as designed".
It's a dumb design since it flies in the face of over 30 years of precedent, and there is already a way to tell if a disk is inserted.
Well, now that I know that a solid light means (essentially) nothing, I will just ignore it. I guess the scrub and defrag were really all I needed, since MySQL and file performance has returned to what I consider to be "acceptable."
I still would like to see the light's function change -- to comply with standards.
There are several standards. On some USB disk you will find a similar working design, the led representing both power and activity. You also have models that have only one led but with colour changing (E.G blue=on, red=activity). I am for standards, that said I don't think that the LED thing is a standard, merely a habit. A standard is meant for compatibility, here that's not the case. I remember times where PC were in plastic cases with only ports on the rear side and green leds. Now we have metal cases with a glass to see inside, lights everywhere, main ports on the front and blue leds. Nobody cared, and some improvements like this were actually useful, like the fact that my Antec P182SE has the power supply at the bottom, it first created problems with cable lengths, but it helped a lot for heat management afterwards.
Microsoft now ships xbox one without kinect, so kinect games won't be developed as intensively as they should have been. I would love to see good kinect games, but with that decision (which was made for Microsoft by folks that refused innovation), I'm not likely to see an hybrid game nor a kinect FPS (remember time crisis 2 in arcade ? I'm pretty sure we can do that with kinect). There is always room to negotiate things and people just tend to accept or refuse, for kinect, a free game (in addition to xbox fitness) or a 1year gold membership or a 2 years warranty extension would have solved the cost issue (and could have been done retroactively for early adopters). We could even imagine that microsoft could have negotiated with health insurance (or state/country to skip sales tax), you get a(n optional) discount on the console (or health insurance fees) but you have to play a kinect game for 2h a week (WHO recommendation) for a year (with a start bonus of 10H to make up for holidays), if you don't get your quota the console enters in limited functionality, preventing non-kinect game launch for exemple. After a year quota checking is disabled. Or a kindle-like model, you get a 30second ad on each boot and each game launch but you get a discount on the console.
I agree that, here, it's mainly cosmetic and comes down to preferences (unless it impacts the LED's life expectancy), but I think it's best to let manufacturers do their things the way they want unless it's clearly impairing the usage, in which case nobody would buy. I believe that innovation begins with creative freedom, if people keep asking for things to look the way they always did, I don't think we will get innovation. As long as it is properly documented I don't see any problem. Maybe they are planning some cheap one bay NAS with only that led ? Not sure how EDA-500 works either, maybe it uses the same system ? Or maybe those leds tend to fail and this is easier to detect it that way ? Of course if this is because an engineer just thought, "hey it's fun !", an option in OS6 would be welcomed, but I don't think they did this without an idea in the back of their head.
There's no industry standard here (certainly not one you can name), but in many cases like this there is industry convergence anyway. The main point 2twisty is making is that it doesn't work the way he expects, which certainly has some weight whether there is a formal standard or not. I don't know if there is a unified industry practice on this or not, but ReadyNAS is the only NAS I've ever owned, and I don't have any data-center experience - so I have no way to judge that.
As far as memories go, personally I used early minicomputers with lots of lights, the ability to enter data (and instructions) into memory from the front panel, hardware breakpoints, etc. When computer front panels disappeared, I missed a lot of those features. The light on the hard drive - if you were lucky enough to have one - meant it had finally finished spin-up and gone on-line so you could use it - which took minutes. It went off when spin-down was finished, so you could remove the platter from the drive - the platters seemed about the size of storm drain covers (well not quite), and placed inside removable cartridges.
BTW, I have participated in many consumer-product evals/betas. On thing everyone has opinions on is the LEDS. Usually its not about what the LED status actually means. Its more about how visible they are. If the product is in your bedroom or home theater, you want the LEDS totally off, and you certainly don't want anything that blinks. If its in an office or rack, you want them bright and easily seen. I totally get this - I have an R7000 in an upstairs bedroom, and I am happy that the GUI has a way to turn the leds off.
I believe I can recognise disk activity when I see and hear it...
I have two ReadyNAS 104 boxes, both with two 4TB drives configured in X-RAID2 as per default. FW version is 6.1.9. One box is just sitting there quietly, showing no significant disk activity unless I access it. The other one is frantically working 24 hours a day. The disk activity LED is blinking all the time and I can hear the drives rattling. What I can't work out is what it's doing. Nobody is accessing the box. A network trace reveals nothing more than an occasional NTP or SSDP packet and such like - nothing that might correlate with disk activity. The disk activity continues even when I pull the network cable. When I reboot the box, things remains quiet for a few minutes after the box comes back up, but then the disk activity resumes. There is no indication of anything happening in the log, except the daily snapshots. I'm not running any extras. I've experimented with backup jobs and cloud backup, but I've turned all of that off.
I don't know exactly when this behaviour started, but it's now been going on for perhaps 10 days. I've had the box only for 2 weeks. In the first 3 or 4 days, I copied 1.7 TB to it. Since then, the box has been idle except for the occasional test. What might explain the contrast with the quiet behaviour of the other ReadyNAS box? This other box has only 500 GB on it, and it's had them for perhaps 2 months.
I don't know if the unexplained disk activity affects performance or has any other unwanted effects, but if it is going to go on, I'm worried first about my electricity bill, and second about the life time of my hard drives.
I was having the exact same issue. Constant disc activity, blinking "act" light, drives making a ton of read/write noise. Restarting ReadyDLNA and disabling autoscan seemed to fix the problem.
I'm seeing constant disk activity too on my new 314. I first noticed it after I installed the Logitech Media Server add-on, so I assumed it was media/metadata scanning, but it's been going on for two full days now. LMS was never switched on (so it can't be scanning); Plex is showing zero scan activity in its logs, and its GUI is often more responsive than the ReadyNAS GUI.
On which point: At first the constant disk activity resulted in an inaccessible ReadyNAS GUI or GUI hangs, and even manually I was unable to power down the 314 gracefully. I had to do a forced power off. Almost immediately after rebooting it started chugging again. This went on for a while, and the GUI was inaccessible, so I tried to do another manual reboot. The "Shutting Down..." message appeared on the LCD — but even after half an hour it hadn't shut down.
That's when I started switching apps and services off, and the unit quieted down for a time. Then it started chugging again, so I added them all back in. I mean, if it's going to chug anyway, I might as well have DLNA access. The only service I've left off is antivirus. Will be downloading my logs and sending them to support (because I can reach the GUI at the moment).
If this is an initial defrag process, I would really, really love to have been notified in some way and to have a progress indicator.
On a side note: So far, I have to say, my experience with the 6.x ReadyNAS unit has been less than stellar. Transfer speeds are far slower than on my 4.x unit, it's less responsive, and apps that were available for the old OS (like the UE Media Server) still aren't available or require hack-y workarounds. The Power Timer interface is just bizarre. I keep hoping things will at least seem to improve as I bring it to the operating state of my Ultra 4.
To follow my previous post, this is what I'm seeing:
The device is up and running and online. My kids were watching a ReadyNAS-hosted movie via Plex and a Mac is in the middle of a Time Machine backup. But there's way more disk activity than there should be, and it's caused the GUI to be totally unresponsive. Never had this issue with the 4.x ReadyNAS.
Do you have SSH port forwarded? I am seeing a lot of login attempts in the logs coming from an I.P. address in China. You might wish to disable port forwarding of SSH to the NAS. There are ways to use e.g. SSH via ReadyNAS Remote
I see you are using snapshots on a lot of shares. These may not be appropriate for all your shares depending on what you use them for. Also on 6.2.0 beta 3 which you are running there is an option to disable COW at a per share level. This may be beneficial for some of your data. Disabling COW only helps for new files added, however.
Snapshot deletion can put a high level of load on the NAS.
It's possible one or more of your add-ons might be preventing the disks from spinning down.
Do you have SSH port forwarded? I am seeing a lot of login attempts in the logs coming from an I.P. address in China.
Thanks, mdgm. I had set up port forwarding ages ago while I was traveling and never disabled it. My logs were full of brute force SSH attempts. (Yeah, I'm looking at you, 76.12.235.157.)
mdgm wrote:
I see you are using snapshots on a lot of shares.
I disabled snapshots as well as bit rot (COW). I might enable them later if I feel like the NAS can handle it. Right now they're not necessary because I still have my Ultra 4 as a backup, and the data hasn't changed greatly.
mdgm wrote:
It's possible one or more of your add-ons might be preventing the disks from spinning down.
My first assumption was Plex, but I think I've found the culprit: iTunes. Something is very, very wrong with forked-daapd based on the logs. When I finally did get the ReadyNAS iTunes library to load in iTunes, this is what I saw:
In case it's not clear, each track has seven entries.
So it looks like the iTunes service scans the media files up to a point, then there's the following series of messages:
Nov 14 21:03:31 ReadyNAS forked-daapd[1907]: main: Got SIGTERM or SIGINT Nov 14 21:03:31 ReadyNAS forked-daapd[1907]: main: Stopping gracefully Nov 14 21:03:31 ReadyNAS forked-daapd[1907]: main: mDNS deinit Nov 14 21:03:31 ReadyNAS forked-daapd[1907]: main: Remote pairing deinit Nov 14 21:03:31 ReadyNAS forked-daapd[1907]: main: HTTPd deinit
Soon forked-daapd enters a "failed state" before restarting. And the cycle repeats itself. So it just keeps scanning and rescanning. Once I disabled iTunes, I had a quiet ReadyNAS again.
Also, it looks like the logs were full of attempts to send alert messages via SMTP every minute or so. I tried setting up Gmail and Zoho accounts, but neither sent successfully. Has anyone successfully configured a mail account to send alert messages? [EDIT: Now mail alerts are working. Go figure.]
