I'm running latest unraid 5 pro, no plugins or addons. To help diagnose the issue and verify it is a stock problem, I did a fresh rebuild of the flash drive and copy over of config directory. The problem did occur with the last beta 16c(?). no router is involved, just two switches (one enterprise, one workgroup) both GB speed.

 

When copying directly to a disk share (no other activities), through ip address (not server name), from local SDD, with directory opus on windows 7, the copy starts, stalls for 20 seconds, then copys a file or two, then gives me a "specified network name is no longer available". The drive being copied to was already spinning.

 

There is a reoccurring error in the syslog that started since the change to the final 5. It looks like the custom shares out of the smb-extra.conf "PUBLIC_READ" are being forced to try all lowercase on the server disk, even though the server share locations do contain capitals. There were no errors like this prior to the update.

 

There is also the "null passwords option is deprecated" from one of my windows xp systems that occurs at times, but did not occur within the time of test error.

 

Nothing in the syslog gives me a hint of the error and shows the share being closed by the client normally, my "main" share is set to high water and a minimum of 31457280 (all of my drives are over 100GB free (least is 4TB and 3% free). The server has 8gb ram and a quad core with dual gb nics (forcedeth driver)

 

I am assuming it has something to do with samba (previous post was ignored and not replied to by anyone). Is there a way to pump smbd logs out at a higher level to see?

 

 

As a side note for the site, is everyone's syslog.txt under 192KB (post limit is terribly small).

 

 

UPDATE: syslog includes samba syslog = 2...I am recreating the issue with no custom smb globals to see if there are any errors.

 

UPDATE: no smb globals, still times out copying. The file in question is a 7GB file. The interesting part is that the file copy is getting to a different point every time.

syslog.txt

There should be earlier logs in /var/log/. Look for syslog.1, syslog.2, etc. Zip them and attach.

These are the logs from 9-19 (the one from the previous post was just the second half of the latest one). The latest clean version (still has the problem), it's syslog only lists the deprecated lines...If you think it is samba, do you want me to create a clean log file of the error with the log level set to all?

 

FYI samba config options that did not change the problem (removed/defaulted in my latest tests)

 

;[global]

;guest ok = no

;workgroup = workgroup

;server string = unraid

;strict allocate = Yes

;host msdfs = No

;msdfs root = No

;socket options = TCP_NODELAY IPTOS_LOWDELAY SO_RCVBUF=65536 SO_SNDBUF=65536

;deadtime = 15

;log level = 2

;syslog = 2

;max log size = 10000

;max connections = 65535

;max open files = 65535

;use sendfile=no

;name resolve order = bcast host lmhosts wins

;security = user

;case sensitive = auto

logs_09-19.zip

The reason I originally only posted the second half of the syslog is because of the less complications.

Previously i was trying to install the latest plex and found that it ran the server into the ground via a memory leak. The problem has been around for some time before that, but after the rebuild, the system is clean and still having the problem. I have been watching the samba log and the syslog for ANYTHING that occurs within the few minutes before the transfer start and a few minutes after the transfer error out, and nothing. Is there any other logging level that i can jack up and reproduce the problem? I know of the samba error logging i can set to all, but i would imagine there is a more advanced logging level for slackware and the unraid base.

 

To upload all of the syslogs, (been running at level 2 for a while) is a couple of meg and probably wouldn't gleen any more info.

the copy starts, stalls for 20 seconds, then copys a file or two, then gives me a "specified network name is no longer available". The drive being copied to was already spinning.

 

I've been having the exact intermittent issue on both 4.6 and 4.7 - started in the last year. I'd sure like to know what's causing it.

 

the copy starts, stalls for 20 seconds, then copys a file or two, then gives me a "specified network name is no longer available". The drive being copied to was already spinning.

 

I've been having the exact intermittent issue on both 4.6 and 4.7 - started in the last year. I'd sure like to know what's causing it.

@flambot:  This post can tell you what fixed that particular issue for me.  I was waiting to see Bsim's full logs before I mention it, because it looks like he has some other serious issues going on there.

 

Thx for the pointer. While I don't want to hijack this thread, I only have 2gb RAM, so that option won't help me. I'd still like to know why the server has been perfect for years, then this suddenly started happening without any changes to the unit.

The server is one of TAM's supermicro servers with 8GB of ram if that helps. Has this serious issue really been unaddressed since 4.x? I am not running on VM.

 

Should i consider dumping the extra 4GB of ram to test it out?

 

How do i tell what actual network card I have (without opening the system) and if there is an issue?

 

I've recently considered going full 64 bit slackware, the latest 14.1 has the latest samba that has quite a few bug fixes. With 8GB of ram, all of the extras shouldn't kill me.

 

logs from the beginning? do you mean the beginning of a server startup?

 

the odd lines could have only come from either the powerdown script or the plex server...both had been removed.

I should be able to give you a clean full syslog from startup to error by tomarrow...The problem is very re creatable.

Sorry for the delay...I've actually been running on the "safe mode" for the last few days. The problem of copying is now intermittent and not easily re creatable. Here is full syslog that might help. Any ideas?

syslog0928.txt

The problem of copying is now intermittent and not easily re creatable.

I see you're still running in 8GB of ram.  Do you have the same intermittent problem when you boot with the mem=4095M option?

 

Sep 26 22:32:02 UnRAID kernel: forcedeth: Reverse Engineered nForce ethernet driver. Version 0.64.

Sep 26 22:32:02 UnRAID kernel: forcedeth 0000:00:08.0: ifname eth0, PHY OUI 0x5043 @ 2, addr 00:30:48:c6:97:52

Sep 26 22:32:02 UnRAID kernel: forcedeth 0000:00:08.0: highdma csum vlan pwrctl mgmt gbit lnktim msi desc-v3

Sep 26 22:32:02 UnRAID kernel: forcedeth 0000:00:09.0: ifname eth1, PHY OUI 0x5043 @ 3, addr 00:30:48:c6:97:53

I see this board has nForce NICs.  Just a shot in the dark here, but do you by any chance have an Intel NIC you could try with?

 

 

this last boot was in safe mode. Doesn't safe mode cut the ram down or do i have to physically pull the ram?

 

I do believe i have an intel nic around, just have to make sure the slots are available.

do i have to physically pull the ram?

don't have to physically pull out anything.  already told you in earlier post you can just add that option in syslinux.cfg and then reboot.

 

 

I do believe i have an intel nic around, just have to make sure the slots are available.

when you try the intel nic remember to disable the onboard nics in the bios

 

 

How full is the target drive? I've seen similar problems when I had less than ~150GB free of 2-3TB drives and then trying to copy or rip blu-ray images (40GB). In my case I could see the target image file size increasing steadily by refreshing a file explorer until it timed out. I've seen somewhere an official recommendation that you should not exceed 95% fill ratio on ReiserFS but cannot remember where. In my case it was not over that limit but I would still see some problems. Perhaps fragmentation could cause similar problems even earlier.

 

Few similar threads:

http://lime-technology.com/forum/index.php?topic=28464.0

http://lime-technology.com/forum/index.php?topic=20757.15

 

do i have to physically pull the ram?

don't have to physically pull out anything.  already told you in earlier post you can just add that option in syslinux.cfg and then reboot.

 

This is a boot option that will appear on an attached monitor every time the system boots. Select the 4Gig option.

do i have to physically pull the ram?

don't have to physically pull out anything.  already told you in earlier post you can just add that option in syslinux.cfg and then reboot.

 

This is a boot option that will appear on an attached monitor every time the system boots.

Not in version 5.0.  He'll have to add it himself, like in this post, and then reboot.

 

 

I just started having the same problem today. I only have 4gb of ram on my server.

I just started having the same problem today.  I only have 4gb of ram on my server.

With 4gb of ram in your server you could still benefit from using that option.  The server behaves differently with it.

 

 

How does the server behave differently with that option? The only time I have heard of anyone using it is when they have more than 4gb of ram.

How does the server behave differently

4095M < 4GB

`free -l`

 

Right I get that 4095mb is less than 4gb, but how does the server behave differently? Its a difference of 1 MB. I guess I just don't understand how removing 1mb from the available memory will make the server do something different? I get when you have 8gb and set it to 4095 it only utilizes essentially half the installed ram. Other than removing that 1mb does the server use different logic or follow a different code path or something?

does the server use different logic or follow a different code path or something?

Yes.  You can try it and see for yourself how it allocates the memory.

 

That fork has two paths:

#1: mem is less than 4gb. (Disables all PAE crap.)

#2: mem is not less than 4gb.  i.e., it is 4gb-or-more.  (Uses signifficant chunks of precious lowmem just to keep track of highmem.)

 

So your 4gb is not a special case -- it falls under #2, just like 8gb or 16gb etc..

 

 

Interesting. I did not know that. I might have to give that a try.

limiting to 4GB does not change anything with the error.

 

append  mem=4095M initrd=bzroot

 

I will try an alternate network card to verify that it is not the card, but i am still leaning towards samba/unraid bug.

 

Also, i've heard of problems that occur when free space becomes lower (i don't quite think i'm low enough with the size drives i have. Does anyone know the symptoms and what are the underlying reasons for the issues?

limiting to 4GB does not change anything with the error.

limiting to 4GB did fix the error for me on the following servers: 

- Supermicro X8SIL-F + Xeon X3440

- Supermicro X7DBE + 2x Xeon L5420

- Supermicro H8DAE + 2x Opteron 275

The problem was reproduceable not only with unRAID, but with other 32-bit distros with PAE.

 

BTW, an alternative way to make the problem go away was booting a 64-bit kernel. (with the same 32-bit userspace)

 

If the mem option doesn't fix it for you, then you must have other problems on top of that.  Your nForce NIC is a prime suspect, as mentioned in earlier post.