rosswaters

Members
  • Posts

    15
  • Joined

  • Last visited

Recent Profile Visitors

The recent visitors block is disabled and is not being shown to other users.

rosswaters's Achievements

Noob

Noob (1/14)

0

Reputation

  1. I am having issues with Serge. I have a very large server that has lots of RAM and CPUs with GPUs in it. I installed serge downloaded 2 models, both did not work. I see them in the directory if I go into the appdata folder though. When I go to the serge GUI it just tells me to download another model and the load model is blank with no models downloaded. I looked at the logs and saw this error. # WARNING Memory overcommit must be enabled! Without it, a background save or replication may fail under low memory condition. Being disabled, it can also cause failures without low memory condition, see https://github.com/jemalloc/jemalloc/issues/1328. To fix this issue add 'vm.overcommit_memory = 1' to /etc/sysctl.conf and then reboot or run the command 'sysctl vm.overcommit_memory=1' for this to take effect. Well that would work great if I'm on Ubuntu or something else with etc/*.conf files. Any thoughts on this. I have way more resources then is required.
  2. Thanks JorgeB for all your help. All outputs were 8 and I am now getting 200MBs per second on all the SAS drives. The backplane on the rear of the server is working at 200 as well so it must be a SAS2. I bought the server used in February and just didn't have the time then to make sure it was configured correctly. When I first got it the server was set for failover, it was configured for a camera surveillance server. While I was able to make UnRAID work with in that configuration it was having issues and was really slow. Now it is much better with speeds I was hoping to achieve. Thanks again for your help. Ross
  3. Thank you JorgeB! The MB is PCI 3.0 and I do have a hand full of 9207's on hand so I'll swap them out. I'm okay with 200MBs a second, 100 is just horrible. The documents says the 12 port backplane is a SAS-826EL2. It has 4 SFF-8087s, Primary 1 & 2 then Secondary 1 & 2. They are all SFF-8087's I have a couple 8088's on my desk and I think I looked at them and then typed out the they were SFF-8088's without thinking about it. All internal ports are 8087s. ( I'm guessing you know that though ) Thanks for the info. I will write this info down in my manual. Ross
  4. I have the Supermicro SC847 Chassis The Backplane is the SAS2-846EL2 From what I can tell you can use 2 SSF-8088 on the front backplane but that only works as failover. It has Primary 1, 2, & 3 and Secondary 1, 2, & 3. Personally I would think all the Primary Ports are for the HBA support with Secondary 1, 2, & 3 for Cascading. The documents are lacking on what each primary SFF-8088 port does and how to use them. For instance nowhere does it talk about Primary 3 and Secondary 3 and how to use them yet they're there on the PCB. Can I connect the 3 Primary Ports up to 3 SFF-8088's and have 300MBs per second on SAS drives? I am using only SAS drives in this unit so there are no SATA's involved. It's a 36 bay server so it has a 24 port backplane on the front and a 12 port on the rear. The Rear has 4 SFF-8088's 2 primary and two secondary's. I currently am using two LSI 9211-8i's HBAs in IT mode but I'm moving to two Areca 1882I SAS controller. I am willing to add a 3rd HBA or Areca controller if I can run 5 SFF-8088's. I am not daisy chaining to another enclosure and I don't need redundancy on the SFF-8088s, I am just looking for better than 100MBs per drive. Thanks for any help. Ross
  5. For "Option 2" What 3 SAS connectors or what one per Backplane do I use. I am having a hard time finding documents on correct performance configuration? For my current setup it is daisy chained and I have twice the drive count with half not existing. I am also getting really bad performance across all drives. I love that it has 36 HDD bays but I with I could just run 9 SSF-8087 cables to them. Thanks for any info you can give, Ross
  6. Thanks Jonathan, I will look into the log. The fact that it failed to read makes me leary, I am going to just toss it as far as being a member of unRAID. I've been using WD 8TB RE drives and Seagate SAS drives. I had this Seagate USB drive that was given to me that was just over two months old and thought I would add it. I will stick with known good drives that are used for NAS hardware and stop playing with desktop versions Thanks again for the steps on removing it. Ross
  7. Hi everyone, I installed a new drive that I had ran a preclear on and checked out fine. After adding it to the array, within a 3 week window it failed. UnRAID now has it set to "Device is Disabled, Contents Emulated". The issue I have at the moment is that the storage pool is now unprotected and the drive that failed is empty, I haven't added any shares to that drive yet. UnRAID failed it when it did a parity check and the drive failed to read. How do I go about in removing it from the array and restore parity to the pool of storage drives with the drive gone, can this be done? again the drive is empty. Thanks for any help on this, Ross
  8. Update, after a few servers that used the ARC-1882IX-12 controllers were taken off line I found that they do in fact work for anyone with the same question. I have yet tested the 24 port cards but it uses the exact driver so I would say yes to that too. You can create pass-through disks and RAID 1-60 arrays and they work. The only issue I found to date is any pass-though disks created will be impossible to delete from the web interface while the server array is online, even if the drive is unassigned. The Areca web interface stated that the drive is in use by the attached system and can't be deleted. Taking array off-line allows pass-through deletion. I now have one of my unRAID servers running one of the 4, 12 port cards I have with plans to add more. I personally run ESX and when vCenter initiates a backup of the VMs I was finding that it was taking an hour to backup my domain controllers alone with the 4TB SAS drive I was using as a cache drive in unRAID. I now have a RAID5 with 4, 1TB SSD drives and I get over 2000MBs until the 4GB ECC memory is full (in my case it takes about 20 seconds at 40G using an Intel XL710-BM2 on both ends) then it drops to 1500MBs ish for the remaining data. The transfers from vCenter now only take a few minutes to perform. I have also decided I am now going to step away from the Windows server environment. I feel this is a good option for me. With super fast cache that can be obtained with hard drives, it make unRAID a viable choice for me since I already own 4, 12 port and 2, 24 port controllers. I still use LSI HBA cards for the rest of the array as they pass information from the drive that is useful to unRAID. For the cache unRAID only sees one drive even though you might have 8 drives making that one drive with redundancy. I am also thinking about making a test server using RAID for both the Cache and the Parity drive to see if I can further boost the performance. I am hoping that the RAID on both the Cache and Parity is enough to allow the server to keep up when all drives are being written to by the users it is supporting. Areca cards pass the temperature of the drives but SMART and other errors have to be set up in the controller using an SMTP server that will pass an email of any events the drives have. They also support SNMP for those with monitoring software for their network devices. This also will increase costs on power usage for home users that try this but I am sure this is minimal when you already have 24-36 drives in the server. I know that 2 exact models with basically the same setup minus the LSI HBAs in unRAID have different power usage over the course of a one month period. The unRAID is about 17% less to operate. I believe this is due to the drives powering down when not in use. The Windows servers use more power cause the drives are on all the time due to the increased time it takes to spin up a 16 disk array. unRAID is set to 1 hour before idle Windows is set to 17 hours Performance match, unRAID wins big time using this configuration in both speed and power usage. I am hoping I can do the same big win when I configure the Parity drive to a RAID array.
  9. Following the advice, I copied files off the USB drive then did a low level format on the drive to be sure there wasn't an issue with it. I then reinstalled unRAID using their image creator, Unraid.USB.Creator.Win32-1.6. After the image was finished I replaced the config folder . Everything booted up and I am back running on the server that was having issues. I would like to know what caused it but at least it was an easy fix. It had been 6 plus months since I last rebooted the server and if I wasn't installing a new Areca RAID card for faster VMs and a bigger cache drive on a RAID5 I wouldn't had even had to reboot it at this moment,
  10. Thanks itimpi, I was guessing that was the case after I looked at the USB structure of my other server. Thanks for the confirmation! I can read and extract all the files with no issues on the USB drive that I thought might be failing. Can I just replace the bzroot file on the USB in question? I will also attach an image of what it stated after being on for 6 hours. It looks like it goes further if you allow it time. Thanks again, Ross
  11. I've been running unRAID for quite some time now and I use them as backup servers for my main storage cluster. I just bought some new 8TB drives and decided to shut the server I had plans to upgrade down so I could verify what drives were in it and where they are located on the backplane. I did however take a screenshot of the drive locations, and their serial numbers so I have the order and what SD? they're associate with, if that helps me. I however failed hard as I don't think I ever backed up the config settings, not sure how but I didn't do it on that server. Jumping ahead 3 days later, I decided to boot the server up to find that the USB drive is haning at bzimage and bzroot. Sometimes it passes the bzimage to bzroot, sometimes it doesn't. It is leading me to think that the Sony USB drive Im using is failing or has failed. Is the config file on the drive, can I just pull it off the drive if I can mount it in Windows? The server last time I checked (Screen capture I took) says 29TBs are written to the drives. It is not a huge deal if I had to format all and start over but that is a lot of data to have to move again. I am running a Supermicro mainboard on that machine and have read that they sometimes have issues with bzroot hanging. I have another unRAID server running a new Asus mainboard and will test the USB in that machine once it is done doing its preclear on all the new 8TB drive I just bought. Thanks for letting me know if the configurations are on the USB and what hey are called. Ross
  12. I have a suggestion, not sure if this is where this goes or if there is another channel for requests. I really like this application as it makes it simple to install items on my server. One suggestion, I would love to see a rating system implemented where others along with myself could rate applications. Maybe even allow for reviews on why someone rated an app low or high. I think this would improve the Community Applications allowing others to see the programs ratings and if there are any issues with the app before installing the it from the interface. Not sure how hard this would be to add, but it'd be a nice added feature on an already awesome easy to use plugin. Thanks, Ross
  13. That is ECC for the cache not ETC, stupid auto correct. I am mostly wondering if the pass through option will work. Thanks again for any info, Ross
  14. Hi, I have two servers that currently have 1 ARC-1882IX-24 controller in each 4U server. Will the controllers work in unRAID. I am currently in testing mode on an older Supermicro Xeon 771 server that I had laying around and am finding some controllers do not work. I really cant take my two data centers off line to test the Areca controllers but I really want to leave the Windows server enviroment. Both have V1.52 2014-02-07 installed and both are the 4GB ETC models. Any info would be great on these cards. Ross