Aigolf Posted March 23, 2013 Share Posted March 23, 2013 here are the results of my parity check within the last 20 minutes with SF 1.11 Simple Features WI Display: 5 mins Estimated speed: 111,1 MB/sec 7 mins Estimated speed: 112,0 MB/sec 8 mins Estimated speed: 108,4 MB/sec 9 mins Estimated speed: 109,3 MB/sec 10 mins Estimated speed: 110,7 MB/sec 11 mins Estimated speed: 111,1 MB/sec 12 mins Estimated speed: 113,0 MB/sec 13 mins Estimated speed: 112,7 MB/sec 14 mins Estimated speed: 109,3 MB/sec 15 mins Estimated speed: 110,7 MB/sec 20 mins Estimated speed: 110,8 MB/sec UNMENU WI Display 20 mins 118.5 MB/s so for me is that ok. althought the paritiy disk is not on the onboard controller because it is used for the esxi data stores. as i said before currently no problems for me with unraid & sf 1.11. cheers Link to comment
jumperalex Posted March 23, 2013 Share Posted March 23, 2013 here are the results of my parity check within the last 20 minutes with SF 1.11 Simple Features WI Display: 5 mins Estimated speed: 111,1 MB/sec 7 mins Estimated speed: 112,0 MB/sec 8 mins Estimated speed: 108,4 MB/sec 9 mins Estimated speed: 109,3 MB/sec 10 mins Estimated speed: 110,7 MB/sec 11 mins Estimated speed: 111,1 MB/sec 12 mins Estimated speed: 113,0 MB/sec 13 mins Estimated speed: 112,7 MB/sec 14 mins Estimated speed: 109,3 MB/sec 15 mins Estimated speed: 110,7 MB/sec 20 mins Estimated speed: 110,8 MB/sec UNMENU WI Display 20 mins 118.5 MB/s so for me is that ok. althought the paritiy disk is not on the onboard controller because it is used for the esxi data stores. as i said before currently no problems for me with unraid & sf 1.11. cheers And what do you get when SF is not installed at all? Link to comment
limetech Posted March 23, 2013 Author Share Posted March 23, 2013 If realtek NIC, from console, type this: lsmod Do you see "r8168" or "r8169" in the list? If not type modprobe r8168 modprobe r8169 lsmod Now is it there? Once one of those shows up in the list type this: /etc/rc.d/rc.inet1 start Then tell me if the network starts working. Nothing changed with the realtek drivers btw vs. -rc11. Link to comment
madburg Posted March 23, 2013 Share Posted March 23, 2013 So no problems with your system, thanks for that. See Dougy? It was usefull. Douchebag. Testing with plugin(s) and posting here is a waste of time. And just pollutes the thread and anyone's understanding of where this RC stands. If the plugin(s) your using, work or not (worse, better, same, etc...) those results can be posted in the plugin forum(s) if you wish to share your results with the author of the plugin and/or anyone else using it. What Tom needs here is unRAID RC12 out of box with no alterations/additions/modifications, etc... to get 5.0 to Final. If you can't or will not test in this matter, post your results in the plugin forum(s). So please stop asking about anyone's result with or without a certain plug here, take it up in the plugin forum(s). It would be good at this point for Tom and/or the mods to start moving these post out of this thread. 5.0 going to Final is not about if a certain plugin works or not in these 5.0 RC's, why is this so hard to understand. Link to comment
Dephcon Posted March 23, 2013 Share Posted March 23, 2013 If realtek NIC, from console, type this: lsmod Do you see "r8168" or "r8169" in the list? If not type modprobe r8168 modprobe r8169 lsmod Now is it there? Once one of those shows up in the list type this: /etc/rc.d/rc.inet1 start Then tell me if the network starts working. Nothing changed with the realtek drivers btw vs. -rc11. Tom, neither r8168 or r8169 are detected via modprobe for Realtek 8111DL Would tossing in a broadcom server NIC do me any good? i think i have a bunch of extras at work. Link to comment
JackBauer Posted March 23, 2013 Share Posted March 23, 2013 5.0 going to Final is not about if a certain plugin works or not in these 5.0 RC's, why is this so hard to understand. +1. Dougy could have been nicer, but his point is valid. I someday look forward to using SF... But right now I just want Unraid 5.0 final, and all this unnecessary plugin discussion here is not helping with getting that out. Link to comment
madburg Posted March 23, 2013 Share Posted March 23, 2013 His point is NOT valid in this thread, it would be valid in the plugin(s) forum. Whats VALID thus far is that there is a realtek nic issue. Link to comment
jowi Posted March 23, 2013 Share Posted March 23, 2013 Since the plugin structure is a feature of unraid, my questions and remarks regarding sf are valid IMHO. The problem could be plugin related, that is why i also post my findings in the SF thread, but it might as well be unraid related. Or both. Link to comment
limetech Posted March 23, 2013 Author Share Posted March 23, 2013 His point is NOT valid in this thread, it would be valid in the plugin(s) forum. Whats VALID thus far is that there is a realtek nic issue. I don't really care about side discussions about SF, not a big deal to me. All data is good data. Obviously would prefer basic testing with 'stock' configuration, but no worries. Link to comment
madburg Posted March 23, 2013 Share Posted March 23, 2013 Since the plugin structure is a feature of unraid Plugin Structure works, add plg, installs plg. Bad plg, bad installation. Missing a promised plugin manager but that is a different story. my questions and remarks regarding sf are valid IMHO. NOT valid here, valid in the plugin(s) forum. The problem could be plugin related, that is why i also post my findings in the SF thread, but it might as well be unraid related. Or both. Might as well be? Wrong, Its ONLY plugin related as (broken record time) tests and posts here should ONLY be with out of box with no alterations/additions/modifications, etc... So since your tested with a plugin installed your test results should ONLY be post to the plugin(s) forum, not here. -Well, missed Tom's post before my reply, he's blessed posting your SF tests here. God bless. Link to comment
madburg Posted March 23, 2013 Share Posted March 23, 2013 Has any change been made to webgui indexer and/or reverting back to logging to the syslog? Any changes/fixes for AFP? Any fix for mover always spinning up the very first disk (#1), when there is no data that is slated for disk#1? -forgot to ask, for slack: 'user' alias; why was this added, I guess maybe an example of what this 'su' alias can be used for (ex. During a plugin installation) or that maybe you started using it internally for something? Bump Link to comment
jowi Posted March 23, 2013 Share Posted March 23, 2013 -Well, missed Tom's post before my reply, he's blessed posting your SF tests here. God bless. I dont understand why you people are so upset about me posting about plugins. The fact that a plugin can negatively influence unraid's functioning is a big deal and should be mentioned in this thread. Maybe there is a way to create some sort of 'protected' plugin structure so a plugin lives in its own space/vm and can never alter the workings of unraid. If something like that is even possible under linux or unraid, i dont know. Link to comment
mikejp Posted March 23, 2013 Share Posted March 23, 2013 If realtek NIC, from console, type this: lsmod Do you see "r8168" or "r8169" in the list? If not type modprobe r8168 modprobe r8169 lsmod Now is it there? Once one of those shows up in the list type this: /etc/rc.d/rc.inet1 start Then tell me if the network starts working. Nothing changed with the realtek drivers btw vs. -rc11. Tom, neither r8168 or r8169 are detected via modprobe for Realtek 8111DL same here with the 8111D Link to comment
madburg Posted March 23, 2013 Share Posted March 23, 2013 -Well, missed Tom's post before my reply, he's blessed posting your SF tests here. God bless. I dont understand why you people are so upset about me posting about plugins. The fact that a plugin can negatively influence unraid's functioning is a big deal and should be mentioned in this thread. Maybe there is a way to create some sort of 'protected' plugin structure so a plugin lives in its own space/vm and can never alter the workings of unraid. If something like that is even possible under linux or unraid, i dont know. Not per say upset. Just your lack of understanding that 5.0 going to Final is not about if a certain plugin works or not in these 5.0 RC's. I purchased a dual unRAID license not a third party plugin. I have been waiting 2 years to get unRAID 5.0 Final, a final for me is one that works as designed and features slated for it. I have AFP issues, so thus far it makes AFP just a teaser. Mover should not wake up disk#1 when either there is no data to move at all or no data slated for that disk. That is a bug. We used to be able to see duplicates in syslog, that has been removed and the indexer is not practical for finding duplicates. A plugin manager was slated for 5.0, havent seen any movement or concrete statement it is on the chopping block for 5.0 A new alias 'user' has been introduced with RC12, would be appreciated that some explanation and/or, documentation around it. I am upgrading at this time to see if anything has changed around AFP, mover and duplicate files, hence those questions to Tom to know what to look out for or not to bother as nothing was changed in those areas. Maybe there is a way to create some sort of 'protected' plugin structure so a plugin lives in its own space/vm and can never alter the workings of unraid. If something like that is even possible under linux or unraid, i dont know. As for this, its like saying why cant Microsoft stop third party developers from creating issues with their applications. If the third party creates a bad driver/installation/ modification to the system, etc... why can this not be stopped, isolated, etc. Ask yourself this, If you wrote a plugin to do something ever minute to every drive, and thus your drives never spin down, what would you like for unRAID the core product to do about that plugin, or that behavior? Link to comment
Aigolf Posted March 23, 2013 Share Posted March 23, 2013 His point is NOT valid in this thread, it would be valid in the plugin(s) forum. Whats VALID thus far is that there is a realtek nic issue. I don't really care about side discussions about SF, not a big deal to me. All data is good data. Obviously would prefer basic testing with 'stock' configuration, but no worries. so... here my results ... with and without SF ... just to be sure... not much difference Unraid RC12/SF 1.11 Unraid RC12 Stock 05 Mins Estimated speed: 111,1 MB/sec Estimated speed: 115.67 MB/sec 10 Mins Estimated speed: 110,7 MB/sec Estimated speed: 113.02 MB/sec 15 Mins Estimated speed: 110,7 MB/sec Estimated speed: 111.89 MB/sec 20 Mins Estimated speed: 110,8 MB/sec Estimated speed: 114.89 MB/sec Cheers Link to comment
madburg Posted March 23, 2013 Share Posted March 23, 2013 Had an unclean shutdown, upon reboot, I selected to start the array in Maintenance Mode. Web GUI states Array started in Maintenance Mode AND started a Parity sync in Maintenance Mode. This should NOT be the case right? Screenshot attached. Link to comment
jowi Posted March 23, 2013 Share Posted March 23, 2013 @madburg, i understand your point, but fact remains, if i run SF 1.0.5 on unraid rc5, there is no change in parity speed, speed initially runs as fast as the disks are, 160MB/s. Whether i watch it under SF or unmenu, doesnt matter. But, if i run SF 1.0.5 on rc12, parity speed starts at 50% or less... and the only thing that changed is... unraid. Link to comment
optiman Posted March 23, 2013 Share Posted March 23, 2013 well I'll wait this one out, given my SuperMicro C2SEE has the onboard Realtek RTL8111C nic. From what has been posted, nobody has RC12 working with this nic. Link to comment
mrow Posted March 23, 2013 Share Posted March 23, 2013 -Well, missed Tom's post before my reply, he's blessed posting your SF tests here. God bless. I dont understand why you people are so upset about me posting about plugins. The fact that a plugin can negatively influence unraid's functioning is a big deal and should be mentioned in this thread. Maybe there is a way to create some sort of 'protected' plugin structure so a plugin lives in its own space/vm and can never alter the workings of unraid. If something like that is even possible under linux or unraid, i dont know. This would probably take quite a bit of work on Tom's part to implement but he could make plugins run in an OS level VM. It's a virtual machine that uses the existing OS kernel as the host OS so it's a lot more lightweight than a traditional VM. It's basically a Linux copy of a feature of feature FreeBSD created a while ago called jails. There are a few ways to implement it in Linux. Feel free to Google it for more info. But running plugins in this type of VM should prevent plugins from affecting the underlying unraid system. This would also mean, however, that a plugin like the SF GUI would not be possible. Link to comment
madburg Posted March 23, 2013 Share Posted March 23, 2013 @madburg, i understand your point, but fact remains, if i run SF 1.0.5 on unraid rc5, there is no change in parity speed, speed initially runs as fast as the disks are, 160MB/s. Whether i watch it under SF or unmenu, doesnt matter. But, if i run SF 1.0.5 on rc12, parity speed starts at 50% or less... and the only thing that changed is... unraid. unRAID is not being written for SF, SF is being written for unRAID. The SF dev's will need to look into what needs to be changed in their plugin. Windows 8 came out, my third party app is not working as it used too in Windows 7 (Or you ran Windows 7 and then installed W7 SP1 and the third party app no longer works as it used too), who do you notify or open the ticket/complaint with? Microsoft or the third party who developed the app, in this case who developed the plugin. IF running 'stock' unRAID RC12 parity checks are fine, and are not fine when running 'non-stock' unRAID then you need to notify the plugin dev's. IF running 'stock' unRAID RC12 parity checks are NOT fine, then you need to notify Tom and supply a syslog. Link to comment
speeding_ant Posted March 23, 2013 Share Posted March 23, 2013 SF will affect the parity check speed while the GUI is open on the main page. This is due to polling the disks for temp, something that Tom has since advised we should not do, so this will be changed in the next release. FYI we only have issues with the temp plugin installed. Regarding this latest RC, unfortunately I'm having the same issue with Realtek NICs. Cheers! Link to comment
limetech Posted March 23, 2013 Author Share Posted March 23, 2013 well I'll wait this one out, given my SuperMicro C2SEE has the onboard Realtek RTL8111C nic. From what has been posted, nobody has RC12 working with this nic. Right, sorry this was a SNAFU on my part... There's a manual step in my release workflow that I sometimes forget and it results in the realtek drivers not getting included in the 'bzimage' file. I'll be posting a -rc12a momentarily... Link to comment
limetech Posted March 23, 2013 Author Share Posted March 23, 2013 -Well, missed Tom's post before my reply, he's blessed posting your SF tests here. God bless. I dont understand why you people are so upset about me posting about plugins. The fact that a plugin can negatively influence unraid's functioning is a big deal and should be mentioned in this thread. Maybe there is a way to create some sort of 'protected' plugin structure so a plugin lives in its own space/vm and can never alter the workings of unraid. If something like that is even possible under linux or unraid, i dont know. No one's mad except madburg.. don't worry about it. Link to comment
limetech Posted March 23, 2013 Author Share Posted March 23, 2013 SF will affect the parity check speed while the GUI is open on the main page. This is due to polling the disks for temp, something that Tom has since advised we should not do, so this will be changed in the next release. FYI we only have issues with the temp plugin installed. Regarding this latest RC, unfortunately I'm having the same issue with Realtek NICs. Cheers! Maybe just don't poll as often. The SMART command is very disruptive to the I/O stream because it flushes everything in process, waits for the flush to finish, then typically the drive heads seek to one end of the drive (because the SMART command is, well, not so smart), updates the SMART data including the temperature. Finally, I/O is permitted to resume. Also the drives will not necessarily receive the SMART command at the same time due to some serialization in the non-read/write I/O flow, and parity will not resume until ALL the drives have processed the SMART command sent to them. Link to comment
madburg Posted March 23, 2013 Share Posted March 23, 2013 -Well, missed Tom's post before my reply, he's blessed posting your SF tests here. God bless. I dont understand why you people are so upset about me posting about plugins. The fact that a plugin can negatively influence unraid's functioning is a big deal and should be mentioned in this thread. Maybe there is a way to create some sort of 'protected' plugin structure so a plugin lives in its own space/vm and can never alter the workings of unraid. If something like that is even possible under linux or unraid, i dont know. No one's mad except madburg.. don't worry about it. Right! Like right now I am Mad that Tom missed a manual step in his release workflow (Not the first time mind you), and even thought it didn't affect me personally But it's understandable when someone works on a product off and on, you forget things.. nothing to worry about. Link to comment
Recommended Posts