unRAID Server Release 5.0-rc6-r8168-test Available


Recommended Posts

I knew it would happen.. was just trying to be helpfull and provide tom with extra logs to go through...

 

Okay, that wasn't clear from your original post.  I had also posted a log previously, with a controlled provocation of the bug, so that my syslog didn't overflow.

 

Anyway, as Tom posted quite recently, the problem has been traced back to a particular commit in the Linux kernel tree and it is hoped that those involved in that will soon address the issue ... there is light at the end of the tunnel!

 

Does this appear to be the last outstanding fault with 5-RC? Other than a little time to check the rest of the issues have been resolved.

Link to comment
  • Replies 257
  • Created
  • Last Reply

Top Posters In This Topic

Throwing my 2 cents in on the parity check speed.

 

I upgraded my backup unraid to umm what is this called? Release 5.0-rc6-r8168-test (that's a mouthful)

On RC4 i was getting a pretty steady 80MB/sec through the first 25% of the check.

on this version, i am bouncing between 55-70 MB/sec. usually closer to 70 through the 25%.

 

Keep in mind this is polluted test. i have several minor unMenu add-on and simple features.

 

Build = 16 2TB data drives.

X9SCM-F, I3-2100, 16GB ECC ram, 2x SASLP-MV8's.

 

I have not tried this on my main Build since it is LSI based.

Link to comment

Been running this for about a day and a half and it's been completely stable. I was forced to run beta 11 previously because I kept having issues running 2x SASLP-MV8s. I've ran a parity check as well without any issues. Will continue to test and update if anything changes.

Link to comment

OK, I may have something:

 

Got RC6-R8168 up and running on a virgin box (exact replica of Raj's Greenleaf-Technology.com's 20-drive budget tower: 2x AOC-SASLP-MV8, Realtek 8111DL) last Wednesday, and I've been copying files to it almost nonstop since.

 

This morning, I finished preclear on 3 new drives.  I powered down the system and moved them from their preclearing slots (I'd spread them out to test my SAS cards, but wanted them "permanently" housed on the top row).

 

When I restarted, I realized that I'd forgotten to check serial numbers, and wanted to make sure I was loading them into "unRAID drives" using the same order that they were physically stacked in the tower, so I pulled each one out (hotswap bays), checked the serial, then plugged them back in to the exact same slots they were in during the boot process.

 

Rather than restarting (hotswap and all), I went straight back to the web interface and tried to add one of the drives as Disk 4.

 

The web interface accepted my selection, and promptly... did nothing (still "unassigned").  Disk 4 is still blank, but now the drive I'd tried to add as Disk 4 no longer appears in my list of available devices.

 

So, being the impatient running-late-for-work guy that I am, I go ahead and try a full power down and start up - and that disk I'd tried to use as disk 4 is STILL missing from the list.  However, the other two drives are still listed, and I just added both to the array without any trouble.

 

Does anyone have any clue whether I'm experiencing a new 5.0 bug, or if this is the sort of thing that happens all the time?  And if it's the latter, does anyone have any idea how to get that "ghost drive" to re-appear in the web interface for inclusion in the array?

 

Thanks!

Link to comment

Bizarre - I think I solved my own problem, but I'm going to leave my post up for posterity unless and until someone indicates this had nothing to do with RC5:

 

(1) Power down, move "ghost drive" to a new slot

(2) Start back up, get the "BIOS Bump" (boot order changed)

(3) Rather than switching BIOS, I powered back down and moved the "ghost drive" back to its trouble slot

(4) Powered back up, launched web interface... and there it is.  Formatting now.

 

I'm guessing this was just some sort of a hardware hiccup, but just in case it's something bigger - I hope this information helps the development process!

Link to comment

My parity speeds seem to have returned to 80-100MB/s after I got every drive to 100+GB free space. Odd how 30GB was enough on earlier versions? This seems strange to me, so i'll report back if I figure out if it was something else that fixed it. I was seeing faster speeds was because I was doing a parity build, not a parity sync. Build went ~100MB/s, then the sync I did afterwards went 40MB/s just like before.

Link to comment

Upgraded from 4.7 to RC6 yesterday, everything went very smoothly. Parity check was OK, started at ~105MB/Sec and finished at ~60MB/Sec, took about 6.5hrs.

 

I was hoping this would fix some dropped packet issues I was having with 4.7 but alas, the issues remain. Doesn't seem to affect anything as they are quite few and far between.

Link to comment

My parity speeds seem to have returned to 80-100MB/s after I got every drive to 100+GB free space. Odd how 30GB was enough on earlier versions? This seems strange to me, so i'll report back if I figure out if it was something else that fixed it.

 

I cannot confirm your findings.  After upgrade from rc5 to rc6 my parity check speed dropped from 80+ to 30.  One of my drives only had 60GB free, the next was 330GB.  Hopeful that you might have a solution, I moved 45GB from that fullest drive, to another drive, leaving 105GB free.  My parity check speed is still stuck at a very solid 32MB/s.

Link to comment

Sad to say one of my SAS2LP machines just got a SAS MV8 error during it's parity rebuild, though it's a bit different than before. Instead of taking all of unRAID down with it, it just red balled the drive and gave write errors. I stopped array, took drive out, put drive back in, did extended SMART test, and verified the drive had 0 SMART errors. Looking at the log, it seems like all 8 drives on the SAS MV8 controller were taken offline and then recovered shortly later, but that was enough to cause the parity rebuild to fail and take the parity drive offline due to write errors.

 

So I'm able to get this to happen at will and in testing with latest kernel (3.4.4) the error recovery still happens' date=' but the extra commands are never "lost" and everything unwinds correctly.  I'm not ready to say this is the solution to the problem, but I need more data points from the user base to determine if it is.[/quote']

 

If I'm reading this right, then that's basically what happened here? The error happened, it recovered itself (I can still access webui, everything else is working, etc). It just didn't recover fast enough during parity rebuild, resulting in write errors.

 

EDIT: Tried to rebuild again, errored again opened new thread: http://lime-technology.com/forum/index.php?topic=21545.0

syslog.txt

Link to comment

Hello;

this afternoon decided to give rc6 a try, after experiencing many times the stale nfs issue.

I've switched from rc5 to rc6-test, and back right away; rc6 proved unusable to me due to the absence of the network, and all the disks (except the flash).

 

I had SimpleFeatures installed in rc5; my 1st attempt was to just replace bzroot and bzimage on the flash drive, leaving everything as it was. It didn't work.

I've renamed then the /config/plugings, /boot/packages and /boot/extra to stop any additional programs from loading. That stopped SimpleFeatures, but the net result was the same: no ethernet driver loaded (no network), and none of the disks was mounted.

 

My server is sort of a relic. May not be of relevance to many - nevertheless, this is my experience with the new release and I'm reporting it.

 

Attaching my  syslogs (both rc5 and rc6) for reference.

 

so.. I'll stay on rc5 for now, until the fate of the new release is decided... Hopefully you'll figure it out sooner than later.

 

Thank you for your hard work.

 

HG

syslog.rc6-test.txt

syslog.rc5.txt

Link to comment

My parity speeds seem to have returned to 80-100MB/s after I got every drive to 100+GB free space. Odd how 30GB was enough on earlier versions? This seems strange to me, so i'll report back if I figure out if it was something else that fixed it.

 

I cannot confirm your findings.  After upgrade from rc5 to rc6 my parity check speed dropped from 80+ to 30.  One of my drives only had 60GB free, the next was 330GB.  Hopeful that you might have a solution, I moved 45GB from that fullest drive, to another drive, leaving 105GB free.  My parity check speed is still stuck at a very solid 32MB/s.

 

After more testing it wasn't the fact I freed space from the drives. I was seeing faster speeds was because I was doing a parity build, not a parity sync. After parity was built, I attempted to sync it again, and I get ~40MB/s. Definitely software related, if anything building should be slower than syncing.

Link to comment

sorry, but just checking in after being gone a few days...and now I see rc6 8168 test.  I'm running rc5.  Is this release for a specific group, or should everyone now be on this new TEST release?

 

My only issue so far is when I reboot.  Sometimes the system halts at boot and the screen shows it was at the point of loading the hd controller.  I have to cycle the power.  It sometimes takes a few attempts and then it will boot normally.  The array never reaches the point to come online so I guess it's not critical in terms of screwing with your data, but it sure sucks.  I will take a screenshot next time it does it to share.  I have the Super C2SEE and a Adaptec RAID 1430SA card.  I'm at 12TB with 8 drives in the array + parity drive + cache drive.  Anyone else have this problem?

Link to comment

sorry, but just checking in after being gone a few days...and now I see rc6 8168 test.  I'm running rc5.  Is this release for a specific group, or should everyone now be on this new TEST release?

 

 

The 8168 test is for people with realtek hardware that doesn't work properly, I think this is the 8169 driver being replaced by the 8168 driver.  You don't need the 8168 release unless you're affected by this particular issue.

Link to comment

sorry, but just checking in after being gone a few days...and now I see rc6 8168 test.  I'm running rc5.  Is this release for a specific group, or should everyone now be on this new TEST release?

 

 

The 8168 test is for people with realtek hardware that doesn't work properly, I think this is the 8169 driver being replaced by the 8168 driver.  You don't need the 8168 release unless you're affected by this particular issue.

 

Thanks, that is what I figured but wanted to confirm.  What is weird is I have a realtek network port on my Super Micro C2SEE motherboard, yet I do not have any network driver issues.  This is what the specs are for my mb,

 

Network Controllers 1 x Realtek RTL8111C Gigabit Ethernet

Supports 10 BASE-T, 100 BASE-TX, and 1000 BASE-T with RJ45 output

 

 

By the way, its mentioned above about a v5 rc6 release.  That release isn't out yet, only rc6 TEST is out re what I see on this forum.  The rest of us are on v5 rc5.

Link to comment
Network Controllers 1 x Realtek RTL8111C Gigabit Ethernet

 

It's the RTL8111E chipset which has been showing problems with the r8169 driver.

 

By the way, its mentioned above about a v5 rc6 release.  That release isn't out yet, only rc6 TEST is out re what I see on this forum.  The rest of us are on v5 rc5.

 

rc6-test is the only one we have (and may be the only rc6 version to see the light of day).

Link to comment

  What is weird is I have a realtek network port on my Super Micro C2SEE motherboard, yet I do not have any network driver issues.

It only showed under some high usage situations.  I have the same C2SEE motherboard on my newer unRAID server, and under a "rsync" from my other older server it would show as a disconnect occasionally.

 

The "rc6-r8168-test" release solved that issue for me.  It has been stable since I installed it, through many rsync cycles.

 

Joe L.

Link to comment

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.