unRAID Server Release 5.0-beta12a Available


limetech

Recommended Posts

Download | Release Notes

 

Note: if you see a crash or some other issue, please include a system log or your post will likely be ignored.

 

Main thing here is continuing the saga of the Realtek drivers.  This is what I have found out so far:

 

Realtek provides drivers for various O.S.'s, including linux.  For our purposes, we are interested in two of them:

a) r8169 - supports NIC chips on PCI bus (but don't confuse with linux r8169, see below), and

b) r8168 - supports NIC chips on PCI-Express bus

 

Linux also provides a "built-in" driver (i.e., one that is part of the kernel distribution), also called "r8169".  This driver historically has supported all Realtek NIC's, those with PCI and PCI-Express interfaces.  Immediately some questions come to mind:

 

1. Why does linux have it's own Realtek driver?  Answer: as near as I can tell there are two reasons.  First, the Realtek drivers specifically include verbiage that there are "Several US Patents covered by this driver".  The linux world is incompatible with patents (that is, they are hated).  Second reason is that kernel development moves much faster than Realtek driver guys.  Result is that as new kernels get released, very often the latest "official" Realtek drivers no longer build correctly.  In general, when "official" Realtek drivers get updated, those changes are rolled into the linux driver (who does this I'm not sure).

 

2. Why is linux 'r8169' named the same as an official Realtek 'r8169' driver?  Probably because the linux version is a fork of the Realtek version, and as Realtek created 'r8168' those changes simply got incorporated into linux 'r8169'.  The choice to keep the name 'r8169' in linux is unfortunate and has led to all kinds of confusion - why (someone) doesn't change it is beyond me to be honest.

 

3. Ok, so what the hell is the problem with linux 'r8169'? Apparently there are fixes that are in unreleased kernel 3.1:

http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=commitdiff;h=70090424e59652c4b2e777b533cc23134b176b83

 

But since current "stable" release is 3.0.4, here is what I'm doing:

 

For this release, unRaid-5.0-beta12a, I once again removed the linux 'r8169' driver, and instead have built BOTH the Realtek r8169 and r8168 drivers.  From scouring the newsgroups it appears these drivers are the most stable at this time.  And lo and behold, they both build correctly with linux 3.0.3 - well after I modified a build file they did...  So I need reports from you guys using Realtek chips - as if I have to ask :)  When linux 3.1 is released, I will produce another release that removes the Realtek drivers and re-instates the linux driver.  Why do I want to do this? Because I'd much rather use the linux version if it works.

 

Another change in this release is a fix to increase a time-out to see if it solves a BLK_EH_NOT_HANDLED issue.  What I can say about this is that apparently another area of development in the 3.0 kernel is to "unify" error handling in the block layer, so probably there are some rough edges here still that should improve over time  ::)

 

Finally, for those of you with huge servers, the number of storage devices recognized is increased to 26 - note there is NO increase in array width yet at this time, this is merely expanding the length of the drop-down box for assign drives on the Main page.

 

Link to comment
  • Replies 383
  • Created
  • Last Reply

Top Posters In This Topic

Top Posters In This Topic

Posted Images

 Thank you for giving us an new beta (B12a) I have downloaded it, but will have to wait til my last drive finishes running preclear, which should be later this evening. I have the flash drive ready to go now.

  I am almost finished building my new server, and want the very latest unRAID OS to run it.

   I look forward to the finished 5.0 release!

Thanks again

       Mike

Link to comment

Hey Tom.

 

Yesterday you were calling this 13?

 

Just wondering why the different name...  Is there a 13 still in development and this is a temporary "patch"?

 

TY :)

 

Yes that's right, I decided to not delay these changes while some other goodies are still being tested for -beta13 (or maybe -rc1 if I can get the Realtek issues and some kernel issues sorted out).

Link to comment

hehe, I'll move to 12a and see if the BLK issue goes away.  I honestly can't say where/why/when it decides to happen.  The box has been up since yesterday afternoon without it appearing in the log.

 

I've also grabbed a 2-ch sata card so I can remove one of the SAS cards and use onboard+2ch.  I'll test beta 12a first though, before I start tearing out parts.

 

 

Link to comment

So I need reports from you guys using Realtek chips - as if I have to ask :)  

 

Just tried beta12a after using beta12.

Unfortunately my write speed has dropped significantly with this release [EDIT: I've checked and my wiring is not at fault].  Now I'm getting a write speed of only 8 MB/s max when transferring a 1.5Gb file.  I'm guessing it's because of the new Realtek driver in this release.

 

My network card is Gigabit capable (and speed was indeed Gigabit on release 12).  My NIC is a PCI Tenda Gigabit TEL9901G which utilises a Realtek r8169 chip.

 

My attempt to resolve this was to try running

ethtool -s eth0 speed 1000

and when that didn't change the speed I tried

 

ethtool -s eth0 speed 1000 duplex full

 

but to no avail: it still shows current speed as "Speed: 100Mb/s" in ethtool even though ethtool indicates that my NIC is indeed capable of 1000Mb/s.

 

syslog & ethtool report attached

I think all the "eth0: link up" and "eth0: link down" messages where about the time when I tried the above "ethtool -s ..." commands.

 

beta12a_syslog_ethtool.zip

Link to comment

So I need reports from you guys using Realtek chips - as if I have to ask :)  

 

Just tried beta12a after using beta12.

Unfortunately my write speed has dropped significantly with this release.  Now I'm getting a write speed of only 8 MB/s max when transferring a 1.5Gb file.  I'm guessing it's because of the new Realtek driver in this release.

 

My network card is Gigabit capable (and speed was indeed Gigabit on release 12).  My NIC is a PCI Tenda Gigabit TEL9901G which utilises a Realtek r8169 chip.

 

My attempt to resolve this was to try running

ethtool -s eth0 speed 1000

and when that didn't change the speed I tried

 

ethtool -s eth0 speed 1000 duplex full

 

but to no avail: it still shows current speed as "Speed: 100Mb/s" in ethtool even though ethtool indicates that my NIC is indeed capable of 1000Mb/s.

 

syslog & ethtool report attached

I think all the "eth0: link up" and "eth0: link down" messages where about the time when I tried the above "ethtool -s ..." commands.

 

 

 

 

Something to try from the console (not telnet), is remove driver and reinstall:

 

rmmod r8169

modprobe r8169

 

Then use ethtool again and see if speed still stuck at 100Mbit/s.

Link to comment

So I need reports from you guys using Realtek chips - as if I have to ask :)  

 

Just tried beta12a after using beta12.

Unfortunately my write speed has dropped significantly with this release.  Now I'm getting a write speed of only 8 MB/s max when transferring a 1.5Gb file.  I'm guessing it's because of the new Realtek driver in this release.

 

My network card is Gigabit capable (and speed was indeed Gigabit on release 12).  My NIC is a PCI Tenda Gigabit TEL9901G which utilises a Realtek r8169 chip.

 

My attempt to resolve this was to try running

ethtool -s eth0 speed 1000

and when that didn't change the speed I tried

 

ethtool -s eth0 speed 1000 duplex full

 

but to no avail: it still shows current speed as "Speed: 100Mb/s" in ethtool even though ethtool indicates that my NIC is indeed capable of 1000Mb/s.

 

syslog & ethtool report attached

I think all the "eth0: link up" and "eth0: link down" messages where about the time when I tried the above "ethtool -s ..." commands.

 

 

 

 

Something to try from the console (not telnet), is remove driver and reinstall:

 

rmmod r8169

modprobe r8169

 

Then use ethtool again and see if speed still stuck at 100Mbit/s.

 

If above gets same result, please try this:

 

rmmod r8169

modprobe r8168

 

And report if the r8168 driver works with this NIC.

Link to comment

Something to try from the console (not telnet), is remove driver and reinstall:

 

rmmod r8169

modprobe r8169

 

Then use ethtool again and see if speed still stuck at 100Mbit/s.

 

If above gets same result, please try this:

 

rmmod r8169

modprobe r8168

 

And report if the r8168 driver works with this NIC.

I tried both of the above suggestions.

The first suggestion resulted in ethtool reporting 1000MB/s speed for a second or so, until reverting back to 100MB/s

The second suggestion: r8168 driver doesn't work with this NIC.

 

Link to comment

Something to try from the console (not telnet), is remove driver and reinstall:

 

rmmod r8169

modprobe r8169

 

Then use ethtool again and see if speed still stuck at 100Mbit/s.

 

If above gets same result, please try this:

 

rmmod r8169

modprobe r8168

 

And report if the r8168 driver works with this NIC.

I tried both of the above suggestions.

The first suggestion resulted in ethtool reporting 1000MB/s speed for a second or so, until reverting back to 100MB/s

The second suggestion: r8168 driver doesn't work with this NIC.

Ok, I've reverted back to beta12 and it seems that I don't have Gigabit on that release either! ???

This leaves me with two options:

1) check wiring (which has not changed since I last noticed I had gigabit)

2) revert to older beta (perhaps beta9) and see if I get gigabit with that driver.

 

 

Link to comment

Upgraded from beta 10 and I get

"menu.c32: not a COM32R image

boot:"

If I press tab and then enter unRAID the system boots but I get the same message on a reboot.

I should add that I wanted to start "fresh" so I used just the beta 12a files and copied over only my config folder.

 

Edit: nevermind. Ran syslinux.exe on the flash and it's fixed.

Link to comment

Greetings

 

While not a specific beta issue may I request Lime-tech include the Bonding module in the next beta. I have several MB's with Realtek Giga NICS but have used Intel NICs and deactivated the Realtek in BIOS.

 

Since I have spare router ports which support link aggregation and a "spare (Realtek) NIC" I would like to try bonding but it is quiet a lot of trouble to compile a single module. It is trivial to adapt the ifenslave package so if the bonding module was included it would make this exercise a lot easier.

 

My current motherboard has a Atheros 1 GIG NIC (atl1c.ko is the module) which despite notes that it can be problematic seems OK on the brief tests I have done.

 

I should add the current motherboard I use has an Atheros NIC (module atl1c.ko) . The Intel NIC card cost more than the whole motherboard so I would like to test if I can get better value if link aggregation helps with performance reliability etc.

 

Regards

 

Ramjet303

Link to comment

I just upgraded from 4.7 to 5.0 Beta 12a and have a question on using it with AFP.

 

The wiki says this:

 

"You must be very careful when enabling AFP export of a disk share when that disk is also enabled for user shares. This is because OS X creates several system directories in the root of the shares, and these directories will show up as user shares. To prevent this, you may exclude the disk(s) from the user share file system on the Share Settings page."

 

I'm not sure I understand this. I have a few user shares I believe. Games (Disk1), Video (Disk2), etc. They currently do not have AFP enabled on it. I currently do not have any excluded disks in the user share file systems that I can see.

 

It says be care when that disk is also enabled for user shares so I think this warning applies to me.

 

I was using unRAID as a SMB drive previously for my Macs and trying to make it look more like a native AFP drive now as well as continue using it with my Windows machines.

 

If I want to add support and make it look like just any other AFP drive, what am I supposed to do? It sounds like I need to stop the user shares and then select the AFP Export to Yes but wouldn't that prevent my Windows machines from working? If I export AFP then don't I trip over that warning?

 

What should my settings be if I want to support both Windows (SMB) and Mac (AFP)?

Link to comment

Greetings

 

While not a specific beta issue may I request Lime-tech include the Bonding module in the next beta. I have several MB's with Realtek Giga NICS but have used Intel NICs and deactivated the Realtek in BIOS.

 

Since I have spare router ports which support link aggregation and a "spare (Realtek) NIC" I would like to try bonding but it is quiet a lot of trouble to compile a single module. It is trivial to adapt the ifenslave package so if the bonding module was included it would make this exercise a lot easier.

 

My current motherboard has a Atheros 1 GIG NIC (atl1c.ko is the module) which despite notes that it can be problematic seems OK on the brief tests I have done.

 

I should add the current motherboard I use has an Atheros NIC (module atl1c.ko) . The Intel NIC card cost more than the whole motherboard so I would like to test if I can get better value if link aggregation helps with performance reliability etc.

 

Regards

 

Ramjet303

 

Is this what you are requesting?

http://www.sgvulcan.com/network-interface-bonding-in-slackware-version-13-1/

 

Link to comment

I just upgraded from 4.7 to 5.0 Beta 12a and have a question on using it with AFP.

 

The wiki says this:

 

"You must be very careful when enabling AFP export of a disk share when that disk is also enabled for user shares. This is because OS X creates several system directories in the root of the shares, and these directories will show up as user shares. To prevent this, you may exclude the disk(s) from the user share file system on the Share Settings page."

 

I'm not sure I understand this. I have a few user shares I believe. Games (Disk1), Video (Disk2), etc. They currently do not have AFP enabled on it. I currently do not have any excluded disks in the user share file systems that I can see.

 

It says be care when that disk is also enabled for user shares so I think this warning applies to me.

 

I was using unRAID as a SMB drive previously for my Macs and trying to make it look more like a native AFP drive now as well as continue using it with my Windows machines.

 

If I want to add support and make it look like just any other AFP drive, what am I supposed to do? It sounds like I need to stop the user shares and then select the AFP Export to Yes but wouldn't that prevent my Windows machines from working? If I export AFP then don't I trip over that warning?

 

What should my settings be if I want to support both Windows (SMB) and Mac (AFP)?

 

In a nutshell, don't export disks over AFP, only export user shares (from the Shares menu). You can export user shares as both SMB and AFP.

Link to comment

Having trouble with AFP Time Machine

 

Created/Cleaned up/Rebooted, Upgraded and still get this error:

 

p  3 21:13:44 stadiumNAS emhttp: shcmd (111): killall -HUP smbd

Sep  3 21:13:44 stadiumNAS emhttp: shcmd (112): ps axc | grep -q rpc.mountd

Sep  3 21:13:44 stadiumNAS emhttp: Restart NFS...

Sep  3 21:13:44 stadiumNAS emhttp: shcmd (113): exportfs -ra |& logger

Sep  3 21:13:44 stadiumNAS emhttp: Restart AFP...

Sep  3 21:13:44 stadiumNAS emhttp: shcmd (114): killall -HUP afpd

Sep  3 21:13:44 stadiumNAS emhttp: shcmd (115): cp /etc/avahi/services/afp.service- /etc/avahi/services/afp.service

Sep  3 21:13:44 stadiumNAS avahi-daemon[5393]: Files changed, reloading.

Sep  3 21:13:44 stadiumNAS avahi-daemon[5393]: Service group file /services/afp.service changed, reloading.

Sep  3 21:13:44 stadiumNAS emhttp: shcmd (116): cp /etc/avahi/services/smb.service- /etc/avahi/services/smb.service

Sep  3 21:13:44 stadiumNAS avahi-daemon[5393]: Files changed, reloading.

Sep  3 21:13:44 stadiumNAS avahi-daemon[5393]: Service group file /services/smb.service changed, reloading.

Sep  3 21:13:44 stadiumNAS emhttp: shcmd (117): /usr/local/sbin/emhttp_event svcs_restarted

Sep  3 21:13:44 stadiumNAS emhttp_event: svcs_restarted

Sep  3 21:13:45 stadiumNAS avahi-daemon[5393]: Service "stadiumNAS" (/services/afp.service) successfully established.

Sep  3 21:13:45 stadiumNAS avahi-daemon[5393]: Service "stadiumNAS-SMB" (/services/smb.service) successfully established.

Sep  3 21:13:45 stadiumNAS kernel: mdcmd (50): spindown 1

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (118): :>/etc/samba/smb-shares.conf

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (119): cp /etc/exports- /etc/exports

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (120): cp /etc/netatalk/AppleVolumes.default- /etc/netatalk/AppleVolumes.default

Sep  3 21:13:54 stadiumNAS shfs/user: duplicate object: /mnt/disk2/./.DS_Store

Sep  3 21:13:54 stadiumNAS shfs/user: duplicate object: /mnt/disk3/./.DS_Store

Sep  3 21:13:54 stadiumNAS shfs/user: duplicate object: /mnt/disk6/./._.DS_Store

Sep  3 21:13:54 stadiumNAS shfs/user: duplicate object: /mnt/disk6/./.DS_Store

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (121): echo '/mnt/user/ISOLibrary -async,no_subtree_check,fsid=100 196.196.196.224/27(rw,no_root_squash)' >>/etc/exports

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (122): echo '/mnt/user/VMDisk2 -async,no_subtree_check,fsid=101 196.196.196.224/27(rw,no_root_squash)' >>/etc/exports

Sep  3 21:13:54 stadiumNAS emhttp: Restart SMB...

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (123): killall -HUP smbd

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (124): ps axc | grep -q rpc.mountd

Sep  3 21:13:54 stadiumNAS emhttp: Restart NFS...

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (125): exportfs -ra |& logger

Sep  3 21:13:54 stadiumNAS emhttp: Restart AFP...

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (126): killall -HUP afpd

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (127): cp /etc/avahi/services/afp.service- /etc/avahi/services/afp.service

Sep  3 21:13:54 stadiumNAS avahi-daemon[5393]: Files changed, reloading.

Sep  3 21:13:54 stadiumNAS avahi-daemon[5393]: Service group file /services/afp.service changed, reloading.

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (128): cp /etc/avahi/services/smb.service- /etc/avahi/services/smb.service

Sep  3 21:13:54 stadiumNAS avahi-daemon[5393]: Files changed, reloading.

Sep  3 21:13:54 stadiumNAS avahi-daemon[5393]: Service group file /services/smb.service changed, reloading.

Sep  3 21:13:54 stadiumNAS emhttp: shcmd (129): /usr/local/sbin/emhttp_event svcs_restarted

Sep  3 21:13:54 stadiumNAS emhttp_event: svcs_restarted

Sep  3 21:13:55 stadiumNAS avahi-daemon[5393]: Service "stadiumNAS" (/services/afp.service) successfully established.

Sep  3 21:13:55 stadiumNAS avahi-daemon[5393]: Service "stadiumNAS-SMB" (/services/smb.service) successfully established.

 

 

Sep  3 21:14:16 stadiumNAS afpd[5697]: dsi_stream_read: len:0, unexpected EOF

Sep  3 21:14:43 stadiumNAS cnid_dbd[5725]: Set syslog logging to level: LOG_NOTE

Sep  3 21:15:00 stadiumNAS afpd[5746]: dsi_stream_read: len:0, unexpected EOF

Link to comment

I just upgraded from 4.7 to 5.0 Beta 12a and have a question on using it with AFP.

 

The wiki says this:

 

"You must be very careful when enabling AFP export of a disk share when that disk is also enabled for user shares. This is because OS X creates several system directories in the root of the shares, and these directories will show up as user shares. To prevent this, you may exclude the disk(s) from the user share file system on the Share Settings page."

 

I'm not sure I understand this. I have a few user shares I believe. Games (Disk1), Video (Disk2), etc. They currently do not have AFP enabled on it. I currently do not have any excluded disks in the user share file systems that I can see.

 

It says be care when that disk is also enabled for user shares so I think this warning applies to me.

 

I was using unRAID as a SMB drive previously for my Macs and trying to make it look more like a native AFP drive now as well as continue using it with my Windows machines.

 

If I want to add support and make it look like just any other AFP drive, what am I supposed to do? It sounds like I need to stop the user shares and then select the AFP Export to Yes but wouldn't that prevent my Windows machines from working? If I export AFP then don't I trip over that warning?

 

What should my settings be if I want to support both Windows (SMB) and Mac (AFP)?

 

In a nutshell, don't export disks over AFP, only export user shares (from the Shares menu). You can export user shares as both SMB and AFP.

 

I see the user shares (from the Shares menu), but I don't see any disk shares. Where do I go to verify I'm not exporting disks over AFP? Has this been removed from 5.0 Beta 12a?

Link to comment

Hi lime Tech (Tom)

 

Re Bonding. Yes the above link does cover what I was referring to. Some years ago I set up link aggregation using Slax and it was fine.

 

All the "extras" rc.bond, ifenslave etc are trivial to set up, its only the bonding module in the library for net drivers that for me at least is a pain to set up since I would need to set up a compiling environment for the UNRAID kernel and driver library. While this is not insurmountable it is a lot to do simply for one module.

 

Perhaps using link aggregation may have no great advantage since disk speeds are the limiter. At present the disks I have (Hitachi 5K3000 2 TB and Seagate @TB Greens start at around 135 MB/sec) and with jperf test I can get network speeds (with tuning )112 MB/Sec). Real world rough tests show around 35-50 MB/SEC using Samba and around 70 -90 MB/SEC using (vs) FTP.

 

I am curious to see if using bonding for 2 GB /sec will yield any improvements. In my case it is effectively free since I have a spare port on the switch and a built in Giga NIC that I do not use over the Intel NIC card I have.

 

While anything that adds unnecessary size to bzroot should be avoided this "extra" make it trivial to test whether there is any real "value added" using bonding.

 

Regards

 

Ramjet303

 

PS I have some appreciation that the beta development is a very long and demanding task but hope this request can be conveniently fitted in.

Link to comment

Also in this version I keep getting the BLK_EH_NOT_HANDLED.

 

Sep  4 09:38:25 Goliath kernel: sas: command 0xee53e300, task 0xf74963c0, timed out: BLK_EH_NOT_HANDLED

Sep  4 09:38:25 Goliath kernel: sas: Enter sas_scsi_recover_host

Sep  4 09:38:25 Goliath kernel: sas: trying to find task 0xf74963c0

Sep  4 09:38:25 Goliath kernel: sas: sas_scsi_find_task: aborting task 0xf74963c0

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 1818:<7>mv_abort_task() mvi=f70e0000 task=f74963c0 slot=f70f15d8 slot_idx=x0

Sep  4 09:38:25 Goliath kernel: sas: sas_scsi_find_task: querying task 0xf74963c0

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 1747:mvs_query_task:rc= 5

Sep  4 09:38:25 Goliath kernel: sas: sas_scsi_find_task: task 0xf74963c0 failed to abort

Sep  4 09:38:25 Goliath kernel: sas: task 0xf74963c0 is not at LU: I_T recover

Sep  4 09:38:25 Goliath kernel: sas: I_T nexus reset for dev 0500000000000000

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 2198:port 5 ctrl sts=0x89800.

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 2200:Port 5 irq sts = 0x1001

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 2226:phy5 Unplug Notice

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 2198:port 5 ctrl sts=0x199800.

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 2200:Port 5 irq sts = 0x1081

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 2198:port 5 ctrl sts=0x199800.

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 2200:Port 5 irq sts = 0x10000

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 2253:notify plug in on phy[5]

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 1338:port 5 attach dev info is 0

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 1340:port 5 attach sas addr is 5

Sep  4 09:38:25 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 379:phy 5 byte dmaded.

Sep  4 09:38:25 Goliath kernel: sas: sas_form_port: phy5 belongs to port1 already(1)!

Sep  4 09:38:28 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 1701:mvs_I_T_nexus_reset for device[1]:rc= 0

Sep  4 09:38:28 Goliath kernel: sas: I_T 0500000000000000 recovered

Sep  4 09:38:28 Goliath kernel: sas: sas_ata_task_done: SAS error 8d

Sep  4 09:38:28 Goliath kernel: ata5: sas eh calling libata port error handler

Sep  4 09:38:28 Goliath kernel: ata6: sas eh calling libata port error handler

Sep  4 09:38:28 Goliath kernel: sas: sas_ata_task_done: SAS error 2

Sep  4 09:38:28 Goliath kernel: ata6: failed to read log page 10h (errno=-5)

Sep  4 09:38:28 Goliath kernel: ata6.00: exception Emask 0x1 SAct 0x1 SErr 0x0 action 0x6 t0

Sep  4 09:38:28 Goliath kernel: ata6.00: failed command: READ FPDMA QUEUED

Sep  4 09:38:28 Goliath kernel: ata6.00: cmd 60/00:00:10:d0:50/02:00:0f:00:00/40 tag 0 ncq 262144 in

Sep  4 09:38:28 Goliath kernel:          res 01/04:04:10:ce:50/00:00:0f:00:00/40 Emask 0x3 (HSM violation)

Sep  4 09:38:28 Goliath kernel: ata6.00: status: { ERR }

Sep  4 09:38:28 Goliath kernel: ata6.00: error: { ABRT }

Sep  4 09:38:28 Goliath kernel: ata6: hard resetting link

Sep  4 09:38:28 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 2198:port 5 ctrl sts=0x89800.

Sep  4 09:38:28 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 2200:Port 5 irq sts = 0x1001001

Sep  4 09:38:28 Goliath kernel: drivers/scsi/mvsas/mv_sas.c 2226:phy5 Unplug Notice

 

etc

 

The length of the drop-down box for assign drives on the Main page is still 20 in my case. Should be 26?

The main page reports beta12a aswell... not sure why I don't see drop-down boxes for 26 drives.. ;-)

 

I only replaced the bzimage and bzroot files, removed the super.dat file (renamed it to super.bak).

After a reboot I reorganized the drives back to their original order. That worked without any problems. Started the parity check again... emhttp stopped responding within 30 mins.  All again due to BLK_EH_NOT_HANDLED.

 

I've got a big system which is ready to go, but failing on this error. Is there a version that works without these issues? I really need to get this started asap. Thanks!

 

 

 

Link to comment

I'm getting higher throughput on my Realtek gigabit NIC with b12a than I did with b12.  With b12, Windows taskmon reported I was peaking around 60% utilization and the file transfer said it was getting around 51MB/sec.  With b12a, it's getting up to 80% utilization (though not as steady as b12, network activity is "sawtoothed") and file transfer speed is 62MB/sec.

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.