bsim Posted October 4, 2013 Author Share Posted October 4, 2013 Not the NIC, ...disabled both nforce nics on board, installed an intel pro gt gigabit card, system came right up without any problems on the network, attempted to copy files and same network name is no longer available. BTW, i am still under the forced 4GB limit. Is there a way to increase the unraid log level to show what is actually going wrong? things i may try... actually removing down to 4gb of ram, installing full slackware 64bit install anything else i can try? Being that i have a pro license, is there any way to escalate my issue? Quote Link to comment
Barziya Posted October 5, 2013 Share Posted October 5, 2013 ...installing full slackware 64bit install... You don't need a full Slackware install for that, just the kernel. You can easily test that theory like this: - Grab the latest prebuilt kernel from the Slackware-64 repo, and the corresponding kernel modules; - Repack unRAID's bzroot to insert the new kernel modules folder in it; - Add an entry in syslinux.cfg that uses the new kernel and the repackaged bzroot; - Edit the go script to NOT start emhttp; - Reboot; - Start Samba manually; - Test to see if you still have that problem or not. If that turns out that this makes your problem go away, then you can easily build yourself a new kernel with the unRAID's driver as module for it. Then you can happily use all the ram you can fit in that server, no need to limit anything. Quote Link to comment
bsim Posted October 5, 2013 Author Share Posted October 5, 2013 Are there any dependencies that unraid uses that are part of the slackware installation that unraid would have issues with (samba, emhttp, specific bins...)? Is there any way to raise the unraid logging level to determine the specific error? Is there any way to escalate my problem? Quote Link to comment
henris Posted October 5, 2013 Share Posted October 5, 2013 Do you have any disk which would have > 1TB free? As commented previously I've seen this very same problem with 2TB disks when free space is in 150-100GB region and on cases even with higher free spaces. So with 4TB disks this could be 300-200GB, I really don't know. Just that we could rule out the free space being the problem. Quote Link to comment
bsim Posted October 5, 2013 Author Share Posted October 5, 2013 is there any logging level that would show that type of error occurring? I understand why the free space is necessary, but the error'ing out occurs on files as small as a 100MB. all 12 of my drives have at least 100GB free, i have it set to save to the drive with the most amount free space first. At this point, I am putting in a mirrored 256 GB cache...i've heard of people saying that a cache drive bypasses a lot of their issues, and the only reason i didn't have a single cache drive before is because of the inherent danger to the most valuable (most recent) files being lost...very unacceptable. After getting to that point, with how much system resources i have and the lack of virtualizing, i will probably go full slackware64 14.1 and add a video/sound card with xbmc....can't wait for xbmc's next release to replace plex with dynamic streaming! For reference, the system motherboard is a SuperMicro H8DME-2. Quote Link to comment
Barziya Posted October 5, 2013 Share Posted October 5, 2013 After getting to that point, with how much system resources i have and the lack of virtualizing, i will probably go full slackware64 14.1 and add a video/sound card with xbmc. Do you mean like abandoning unRAID? Or, do you mean making all that work in unRAID? Either way, good luck with that. Let us know how it goes. Quote Link to comment
vl1969 Posted October 6, 2013 Share Posted October 6, 2013 I have the same sm mb. I got it with my TAMs server. And you can virtualize on it with a right cpu. And using xen on ubuntu 12.4 With a litle elbow grees it possible. There is a thread from ddeeds in virtualization section for refference. But basically Go through the bios and make sure to enable all references to iommu and virtualisation. Get a Ubuntu 12.4 lsi server. Install and update. Configure xen 4.2 and networking. Download the xen enabled unraid immage from one of the threads. It is the same for xen and xenserver. Follow the directions closely. The most difficult is figuring out ids for usb and pci controllers. Onec you get that right it works good. I am still waiting for guide on how to install archipel as xen on ubuntu is cli only deal. And not very pretty. But works. Sent from my SGH-T889 using Tapatalk 4 Quote Link to comment
henris Posted October 6, 2013 Share Posted October 6, 2013 is there any logging level that would show that type of error occurring? I understand why the free space is necessary, but the error'ing out occurs on files as small as a 100MB. all 12 of my drives have at least 100GB free, i have it set to save to the drive with the most amount free space first. If you have only 100GB free space on all of your 2TB or larger drives, this is probably the source of the problem. Couldn't you add the 256GB drive as a data disk? Quote Link to comment
bsim Posted October 7, 2013 Author Share Posted October 7, 2013 vl1969: The motherboard does support some virtualization features, but for my sata cards to be directly assigned to a virtual machine, the function isn't supported. I had intended to run vmware esxi 5.1. Accessing pci devices with xen, is this a shared resource or is it dedicated access (i would think that this would have to be a shared). Barziya: From what i've read, Unraid can be ran on top of a full slackware 64bit installation. There are quite a few features that I would find useful that aren't in the appliance-like unraid build. henris: Is there a way to directly determine if this is the cause? The problem did occur before the space got this low (i've been doing some direct drive to drive transfers). The lowest is 100GB, the largest is around 250GB. The files that can cause the error can be as low as 100'M'B. Quote Link to comment
Barziya Posted October 7, 2013 Share Posted October 7, 2013 ...full slackware 64bit installation.. There are quite a few features that I would find useful that aren't in the appliance-like unraid build. Right. That's not the subject of this thread though. We were trying to trouble shoot a particular problem. Quote Link to comment
bsim Posted October 7, 2013 Author Share Posted October 7, 2013 i'm only thinking out loud about the options that solve my end problem. I would like to solve the issue with unraid, but without any other options to tell what the problem actually is, I'm looking at the bigger picture. Is there anything else i can do? Quote Link to comment
vl1969 Posted October 7, 2013 Share Posted October 7, 2013 vl1969: The motherboard does support some virtualization features, but for my sata cards to be directly assigned to a virtual machine, the function isn't supported. I had intended to run vmware esxi 5.1. Accessing pci devices with xen, is this a shared resource or is it dedicated access (i would think that this would have to be a shared). Barziya: From what i've read, Unraid can be ran on top of a full slackware 64bit installation. There are quite a few features that I would find useful that aren't in the appliance-like unraid build. henris: Is there a way to directly determine if this is the cause? The problem did occur before the space got this low (i've been doing some direct drive to drive transfers). The lowest is 100GB, the largest is around 250GB. The files that can cause the error can be as low as 100'M'B. I have tried almost all vm hipervisors , exept kvm and hyper-v. Only xen on ubuntu 12.4 dom0 worked. As for pci access, I am not sure what you asked there. When you do pci pass through to a vm it should be a dedicatd resource, nothing else can access it, not even dom0. Also, I am not sure but I neve heard pci cards that prevent pass through. Isn't that what pass through is all about? Using hardware that not supported by host but suppoeted by vm os? Sent from my SGH-T889 using Tapatalk 4 Quote Link to comment
henris Posted October 7, 2013 Share Posted October 7, 2013 henris: Is there a way to directly determine if this is the cause? The problem did occur before the space got this low (i've been doing some direct drive to drive transfers). The lowest is 100GB, the largest is around 250GB. The files that can cause the error can be as low as 100'M'B. I've seen this happen with a 50kB subtitle file when the free space was low (~100GB) but it is more frequent with larger files. I would just put a new disk in the system (size does not matter) and do testing with it. I would start copying stuff in separated batches and see when it starts to fail. If it starts to fail only after free space is lower than 250GB-100GB then that's the cause. I haven't seen any other debugging method in these forums for this problem. Quote Link to comment
bsim Posted October 7, 2013 Author Share Posted October 7, 2013 vl1969: The motherboard does support some virtualization features, but for my sata cards to be directly assigned to a virtual machine, the function isn't supported. I had intended to run vmware esxi 5.1. Accessing pci devices with xen, is this a shared resource or is it dedicated access (i would think that this would have to be a shared). Barziya: From what i've read, Unraid can be ran on top of a full slackware 64bit installation. There are quite a few features that I would find useful that aren't in the appliance-like unraid build. henris: Is there a way to directly determine if this is the cause? The problem did occur before the space got this low (i've been doing some direct drive to drive transfers). The lowest is 100GB, the largest is around 250GB. The files that can cause the error can be as low as 100'M'B. I have tried almost all vm hipervisors , exept kvm and hyper-v. Only xen on ubuntu 12.4 dom0 worked. As for pci access, I am not sure what you asked there. When you do pci pass through to a vm it should be a dedicatd resource, nothing else can access it, not even dom0. Also, I am not sure but I neve heard pci cards that prevent pass through. Isn't that what pass through is all about? Using hardware that not supported by host but supported by vm os? Sent from my SGH-T889 using Tapatalk 4 I took a week researching the motherboard, and a few dozen posts to an esxi forum and found that the motherboard was made at a time that virtualization was still being standardized and the mb's bios virtualization features are a bit half baked. xen must use a workaround for hardware dedication. There is a way to force the hardware virtualization work around in esxi as well, but there were quite a few drawbacks for doing it so I ended up giving up. The site that explains the subject with xen and the mb's half baked AMD-V... http://en.wikipedia.org/wiki/Hypervisor Quote Link to comment
vl1969 Posted October 7, 2013 Share Posted October 7, 2013 vl1969: The motherboard does support some virtualization features, but for my sata cards to be directly assigned to a virtual machine, the function isn't supported. I had intended to run vmware esxi 5.1. Accessing pci devices with xen, is this a shared resource or is it dedicated access (i would think that this would have to be a shared). Barziya: From what i've read, Unraid can be ran on top of a full slackware 64bit installation. There are quite a few features that I would find useful that aren't in the appliance-like unraid build. henris: Is there a way to directly determine if this is the cause? The problem did occur before the space got this low (i've been doing some direct drive to drive transfers). The lowest is 100GB, the largest is around 250GB. The files that can cause the error can be as low as 100'M'B. I have tried almost all vm hipervisors , exept kvm and hyper-v. Only xen on ubuntu 12.4 dom0 worked. As for pci access, I am not sure what you asked there. When you do pci pass through to a vm it should be a dedicatd resource, nothing else can access it, not even dom0. Also, I am not sure but I neve heard pci cards that prevent pass through. Isn't that what pass through is all about? Using hardware that not supported by host but supported by vm os? Sent from my SGH-T889 using Tapatalk 4 I took a week researching the motherboard, and a few dozen posts to an esxi forum and found that the motherboard was made at a time that virtualization was still being standardized and the mb's bios virtualization features are a bit half baked. xen must use a workaround for hardware dedication. There is a way to force the hardware virtualization work around in esxi as well, but there were quite a few drawbacks for doing it so I ended up giving up. I am not sure how Xen does what it dose :-) but I have found that Xen is the only HV that works with this MB. not easy, and not very intuitive to setup but works. also I could only set it up on Ubuntu 12.4 and Xen 4.2, so far. my try on Ubuntu 13.x and Xen 4.3 failed even though I folow the same procedure in both. even XenServer/XCP did not work for unRaid (pci passthough). ESXi blatantly said that hardware is incompartible with direct access :-) but since I just got the server and have no budget to swap the whole thing into something compatible I have to work with what I have. Quote Link to comment
Barziya Posted October 7, 2013 Share Posted October 7, 2013 Is there anything else i can do? A few things come to mind. Try this: sysctl -w net.core.rmem_default=163840 sysctl -w net.core.wmem_default=163840 sysctl -w net.core.rmem_max=16777216 sysctl -w net.core.wmem_max=16777216 sysctl -w net.ipv4.tcp_rmem='4096 131072 16777216' sysctl -w net.ipv4.tcp_wmem='4096 131072 16777216' sysctl -w net.ipv4.tcp_window_scaling=1 sysctl -w net.ipv4.tcp_mtu_probing=1 ifconfig eth0 txqueuelen 3000 ifconfig eth1 txqueuelen 3000 sysctl -w net.core.netdev_max_backlog=3000 sysctl -w net.ipv4.tcp_max_syn_backlog=4096 sysctl -w vm.min_free_kbytes=65536 read xx MEM xx <<<`grep 'MemTotal:' /proc/meminfo` if [ "$MEM" -gt 2048000 ] ;then sysctl -w vm.min_free_kbytes=131072 ;fi With the above tweaks applied, can you still reproduce the "network name is no longer available" problem? Quote Link to comment
bsim Posted October 17, 2013 Author Share Posted October 17, 2013 just got the last post, I will give it a try. In the meantime, I added a 250GB cache drive directly connected to the MB sata, and freed up almost a terabyte over the array. Even with the cache drive (empty), the system still gives me the error. I will give the tweaks a try to see if i can reproduce the issue. Quote Link to comment
bsim Posted October 17, 2013 Author Share Posted October 17, 2013 UPDATE: The tweaks didn't do anything for the problem. Would the drives being not insanely loaded with free space still be an issue with a cache drive? Attached is an image of my free space situation if anyone was interested. Is there anything else I can do? UPDATE: I found that moving large amounts of files around on the same share can cause the issue. Quote Link to comment
henris Posted October 17, 2013 Share Posted October 17, 2013 I've seen the problem appear on 1-2TB disks when the free space was in 100-150GB region. Unfortunately I do not have cache disk on my production system, where I have had these issues, so I cannot say for certain whether the data disk fill ratio could affect also cache disk writes. To me it would seem odd if it did since they should be totally isolated until the mover kicks in. Could you temporarily re-arrange the data on the disks so that you had one disk with >300GB free (and others with less than 100GB) and test with it? You could/should use a telnet connection do to this since it might not work over smb. This is the last thing which comes to mind apart from getting a completely new data disk. Quote Link to comment
bsim Posted November 13, 2013 Author Share Posted November 13, 2013 After doing some rearranging and a adding a few drives (for space), the issue still occurs. Free space is not the problem with how much space is available on both the cache drive and a few of the drives in the array. Attached is a current drive listing of my server. I do notice that just before the error occurs when copying files to the server, any reads (playing movie on local system) locks until the error fully times out. I'm still running a 5.0 stock 4GB memory limit, the error occurs without syslog events, the network card had been swapped for an Intel card, still no go. I'm still leaning towards a samba/unraid bug. Is there anything more that I can do to diagnose this? Is there a way to increase logging levels in unraid? Quote Link to comment
hpad06 Posted December 3, 2016 Share Posted December 3, 2016 Is there a solution for v5 ? Do you have the same issue with v6 ? I am still running v5 basic version, it is bothering me a lot, would upgrade to v6 fix it ? Quote Link to comment
JonathanM Posted December 3, 2016 Share Posted December 3, 2016 Is there a solution for v5 ? Do you have the same issue with v6 ? I am still running v5 basic version, it is bothering me a lot, would upgrade to v6 fix it ? If your drives are actively written to and some files erased regularly and are rather full, then V6 will probably be worse, not better. Most of the time the timeouts are caused by the delay in allocating open spaces to write new files. Upgrading to V6 allows you to reformat to XFS, which seems to be a little better about handling large full drives than ReiserFS is. However... you would probably solve your issue by reformatting and copying your data back to freshly formatted drives anyway, even if you stay with V5 and ReiserFS. V5 is WAY out of date, both features and security, so you really should think about updating regardless. Quote Link to comment
Recommended Posts
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.