General discussion moved from another thread!!

    • flmaxey wrote:

      I have a post "pending" for some reason, that shows a screen shot of share permissions. It might be awhile before the moderators clear it

      geaves wrote:

      We must have reached or posting limit ....I seem to get errors regularly...just refresh the page and carry on
      The longer the post, the more it has a chance to flagged as spam especially if it has links in it.
      omv 4.0.5 arrakis | 64 bit | 4.12 backports kernel | omvextrasorg 4.0.2
      omv-extras.org plugins source code and issue tracker - github.com/OpenMediaVault-Plugin-Developers

      Please don't PM for support... Too many PMs!
    • ryecoaaron wrote:

      The longer the post, the more it has a chance to flagged as spam especially if it has links in it.
      I never thought about it that way but, from your perspective, stuff by a NOOB might seem like SPAM :D .
      Good backup takes the "drama" out of computing
      ____________________________________
      OMV 3.0.81 Erasmus
      ThinkServer TS140, 12GB ECC /
      32GB USB3.0 / 4TB SG / 1TB WD

      OMV 3.0.81 Erasmus - Rsync'ed Backup Server
      R-PI 2 $29 / 16GB SD Card $8 / Real Time Clock $1.86
      4TB WD My Passport $119
    • There was something else I was going to ask in the main forum....under file system when you plug in a usb you can mount it from the gui, but you can't unmount it as the 'button' is greyed out. Ok this can be via ssh not a problem, is that the same on yours?

      As this Pi is simply going to be used for rsync purposes I could just shut it down with the usb connected and switch it back on when needed, rather than leave it on.
    • geaves wrote:

      1. There was something else I was going to ask in the main forum....under file system when you plug in a usb you can mount it from the gui, but you can't unmount it as the 'button' is greyed out. Ok this can be via ssh not a problem, is that the same on yours?

      2. As this Pi is simply going to be used for rsync purposes I could just shut it down with the usb connected and switch it back on when needed, rather than leave it on.
      1. In my case, since it's the data drive and it's USB powered, my USB drive is always plugged into the R-PI. I've never had a reason to disconnect it.

      If the "unmount" button it greyed out, it's because the drive is referenced in some manner. The openmediavault-resetperms 3.3 plugin adds a tab called "Shared Folder In Use" to Shared Folders. It shows if a folder is in use and what is referencing it. (It's a small and useful plugin. Check it out.)
      If a shared folder is in use and that shared folder exists on your USB drive, the USB drive is referenced and is, therefore, "in use". Hence, the GUI won't let you unmount the drive. ((It makes sense really. If one could do these things, willy nilly, questions along these lines would be eating up the forum.))

      If you know what you want to do, and have a command line that will do it, you could create a Scheduled Job (really a cron job), click on the job in the GUI, and Run it manually with a couple clicks of the mouse. This would do away with having to SSH in and hand typing the command line, every time.

      2. Where the latter is concerned, I actually have my R-PI on an AC timer. This is what I'm using -> Switch. (It comes in a 220V model as well.) R-PI's boot when power is applied so the switch powers on, on Sunday at 0001. I let it boot and give it a 1/2 hour or so. Then I have a series of Rsync jobs run, throughout the day (approximately 1 hour apart), to sync everything up. In the afternoon, I have a cron job that runs a BTRFS "scub". A few hours later, when I know the scrub is complete, I set a cron job to collect the stat's from the scrub and E-mail them to me. A last cron job shuts the R-PI down at 2300. The programmable switch turns off on Sunday at 2359.
      _____________________________________
      The reason I do this, BTW, is not to save power. It's to reduce spinning time on the hard drive. With a drive that may last 3 to 5 years in 24x7 op's, if it's on 1/7th of the time it's life increases substantially (to 21 to 35 years). This also has the added effect of reducing exposure to voltages surges, dips, spikes, etc.
      _____________________________________

      In your case;
      If you have a drive dock, yes, you can shut off the R-PI and leave the dock on. I've done this with an R-PI and a dock before.
      I also believe that you could put a drive dock and the R-PI on a switch like the above, and power everything on and off, at the same time. I've done the same thing, more or less, manually.

      In my case, with a USB powered 2.5" drive, cutting power to the R-PI cuts power to the drive. Powering on powers up both at the same time and all works fine.
      Good backup takes the "drama" out of computing
      ____________________________________
      OMV 3.0.81 Erasmus
      ThinkServer TS140, 12GB ECC /
      32GB USB3.0 / 4TB SG / 1TB WD

      OMV 3.0.81 Erasmus - Rsync'ed Backup Server
      R-PI 2 $29 / 16GB SD Card $8 / Real Time Clock $1.86
      4TB WD My Passport $119

      The post was edited 1 time, last by flmaxey: edits ().

    • flmaxey wrote:

      With a drive that may last 3 to 5 years in 24x7 op's, if it's on 1/7th of the time it's life increases substantially (to 21 to 35 years)
      I disagree with this. In my opinion, the spinup is harder on the drive than running non stop for a long period of time. With a good drive, if you start it 20+ times a day, I bet it doesn't live longer than a drive that ran non-stop for the same total period of time. I realize it uses more energy but I think it is safer when it comes to someone's data.
      omv 4.0.5 arrakis | 64 bit | 4.12 backports kernel | omvextrasorg 4.0.2
      omv-extras.org plugins source code and issue tracker - github.com/OpenMediaVault-Plugin-Developers

      Please don't PM for support... Too many PMs!
    • ryecoaaron wrote:

      flmaxey wrote:

      With a drive that may last 3 to 5 years in 24x7 op's, if it's on 1/7th of the time it's life increases substantially (to 21 to 35 years)
      I disagree with this. In my opinion, the spinup is harder on the drive than running non stop for a long period of time. With a good drive, if you start it 20+ times a day, I bet it doesn't live longer than a drive that ran non-stop for the same total period of time. I realize it uses more energy but I think it is safer when it comes to someone's data.
      But we're not talking about starting a drive 20 times a day. For preservation purposes, I'm talking about starting a drive once a week, or 52 times a year total. On the other side of that particular coin, I have a HTPC and a workstation that I start, pretty much every day, they tend to be up 12+ hours a day, and they've been going through these cycles for more than two years without incident (so far). So, if it's about start cycles alone; stacking up 52 starts against 300+ starts in a year; if you were betting on a failure, which way would you go?

      Also, I believe there are two parts to the drive preservation equation. Starts and spinning hours. Both must be considered.

      In the second half of the equation, I also believe drives last substantially longer if their spinning hours are low.
      Given that consideration, if two drives are started 1 time per week:
      Drive 1 is started, runs for 24 hours, is shut down and cycled on 6 days later.
      Drive 2 is started, runs 168 hours and is power cycled at the end of the week.
      Which drive lasts longer?

      I would agree that 21 to 35 years of drive life may be optimistic but that was simple math, to illustrate a point, without getting into the weeds. However real world observation shows that drive life can be extended to at least 10 to 11+ years, and that particular drive experiment is ongoing. In that particular case, the drives are active once every three months or so, for a few hours, which is enough time to replicate data changes and to stir the lubricant on their spindle bearings (which, I believe, helps to keep them from drying out). Also, I keep these drives in climate controlled conditions where temperatures and humidity don't vary widely or rapidly.

      In the bottom line:
      If you have drives that have been running 24x7 for over 10 years, that are still viable, I'll tip my hat to you. (And I'd appreciate a few lines about how you do it.)
      Otherwise, I must be on to something and it appears that I'm doing it right.
      Good backup takes the "drama" out of computing
      ____________________________________
      OMV 3.0.81 Erasmus
      ThinkServer TS140, 12GB ECC /
      32GB USB3.0 / 4TB SG / 1TB WD

      OMV 3.0.81 Erasmus - Rsync'ed Backup Server
      R-PI 2 $29 / 16GB SD Card $8 / Real Time Clock $1.86
      4TB WD My Passport $119

      The post was edited 1 time, last by flmaxey: e ().

    • flmaxey wrote:

      If you have a drive dock, yes, you can shut off the R-PI and leave the dock on. I've done this with an R-PI and a dock before.
      It's not a drive dock, it's a powered external usb drive, so in essence I can move it around if needs be. I have been looking at a 2.5" WD Red NAS in it's own external case and just powering it directly off the Pi, then just shutdown and restart the Pi as needed.


      flmaxey wrote:

      If the "unmount" button it greyed out, it's because the drive is referenced in some manner.
      That makes sense, but never knew that.


      flmaxey wrote:

      Where the latter is concerned, I actually have my R-PI on an AC timer. This is what I'm using -> Switch. (It comes in a 220V model as well.
      We have a few of those around, used for turning lamps on and off when we are away.
    • flmaxey wrote:

      ryecoaaron wrote:

      flmaxey wrote:

      With a drive that may last 3 to 5 years in 24x7 op's, if it's on 1/7th of the time it's life increases substantially (to 21 to 35 years)
      I disagree with this. In my opinion, the spinup is harder on the drive than running non stop for a long period of time. With a good drive, if you start it 20+ times a day, I bet it doesn't live longer than a drive that ran non-stop for the same total period of time. I realize it uses more energy but I think it is safer when it comes to someone's data.
      But we're not talking about starting a drive 20 times a day. For preservation purposes, I'm talking about starting a drive once a week, or 52 times a year total. On the other side of that particular coin, I have a HTPC and a workstation that I start, pretty much every day, they tend to be up 12+ hours a day, and they've been going through these cycles for more than two years without incident (so far). So, if it's about start cycles alone; stacking up 52 starts against 300+ starts in a year; if you were betting on a failure, which way would you go?
      Also, I believe there are two parts to the drive preservation equation. Starts and spinning hours. Both must be considered.

      In the second half of the equation, I also believe drives last substantially longer if their spinning hours are low.
      Given that consideration, if two drives are started 1 time per week:
      Drive 1 is started, runs for 24 hours, is shut down and cycled on 6 days later.
      Drive 2 is started, runs 168 hours and is power cycled at the end of the week.
      Which drive lasts longer?

      I would agree that 21 to 35 years of drive life may be optimistic but that was simple math, to illustrate a point, without getting into the weeds. However real world observation shows that drive life can be extended to at least 10 to 11+ years, and that particular drive experiment is ongoing. In that particular case, the drives are active once every three months or so, for a few hours, which is enough time to replicate data changes and to stir the lubricant on their spindle bearings (which, I believe, helps to keep them from drying out). Also, I keep these drives in climate controlled conditions where temperatures and humidity don't vary widely or rapidly.

      In the bottom line:
      If you have drives that have been running 24x7 for over 10 years, that are still viable, I'll tip my hat to you. (And I'd appreciate a few lines about how you do it.)
      Otherwise, I must be on to something and it appears that I'm doing it right.
      Well I would agree and disagree with both, I've always looked at it from two angles, if it's a server leave it one, if it's a workstation turn it off. In the case of the Pi as this is going to do a specific job i.e. rsync I would probably only run once a week so in that case it would be powered on as and when required.

      I am sure the knowledge between the two of you far outweighs my own, but I've always worked on two principles; if ain't broke don't fix it, if it's not needed turn it off.

      I have dealt with two IT companies over the last few years, one advocated shut everything down other than the server, the second suggested that workstations and the server should be left but the monitor switched off.

      I would agree that powering a hard drive unnecessarily on and off during a day is going to reduce it's lifespan, but technology has moved forward so much since the early days that pc components are far more robust. I have a 3.5" Maxtor 80Gb which I replaced for a friend, the date on the drive is 2001 but it still works, albeit it sounds like a bag of bolts when it's use.

      What I have a tendency to do today is too look at a drives warranty, with most drives having 3 years and some 5, I then equate that back to 'cost per week'.

      TBH I think you have to look at requirements and the scenerio....if I look at myself for instance, before omv I had a headless ubuntu server which was predominantly for a media server, but it was becoming a 'storage server', old files etc you may need again. Then I had another server running nas4free which was basically my media storage, when I discovered omv and this forum I re evaluated the hardware and with some simple upgrades I now have 1 server that does the job of 2, but in essence actually does more as it runs a single vm as well.

      This is a debate that I find people will agree to disagree, I mean we don't all go for a morning run and come back and have a cold just because we find it invigorating :D :D
    • flmaxey wrote:

      But we're not talking about starting a drive 20 times a day. For preservation purposes, I'm talking about starting a drive once a week, or 52 times a year total. On the other side of that particular coin, I have a HTPC and a workstation that I start, pretty much every day, they tend to be up 12+ hours a day, and they've been going through these cycles for more than two years without incident (so far). So, if it's about start cycles alone; stacking up 52 starts against 300+ starts in a year; if you were betting on a failure, which way would you go?
      Most people aren't on this forum aren't starting a drive once a week though. People want drives to go to sleep when inactive. So, I see drives spinning up and down many times per day. I would still take drives that have 2 starts per year running 24/7/365 over drives with 52 starts per year. Read my next response as to why.

      flmaxey wrote:

      If you have drives that have been running 24x7 for over 10 years, that are still viable, I'll tip my hat to you. (And I'd appreciate a few lines about how you do it.)
      The drives (8 x 2TB Samsung F4) I took out of my server about six months ago were 6 years old and run 24x7 in a raid 5 array. When I took them out, the start stop count was 8 according to smartctl. I was still having zero issues with the drives. I just needed more space and replaced them with WD Red 4TB. The F4s were in my server in my basement with lots of air flow and average temp was about 20 C. Do I think they would last 10 years? Yep. Would I use them that long? Nope. I generally replace drives after 5 years. In my opinion, keeping the drive cool and start-stops low is what has worked for me. I don't remember the last time I had a server drive fail other than my 17 year old 9 GB Dell scsi drive that hadn't been started in 8 years :) I do still have a 200 MB laptop hard drive from my laptop I got in 1992 that still works.

      flmaxey wrote:

      Drive 1 is started, runs for 24 hours, is shut down and cycled on 6 days later.
      Drive 2 is started, runs 168 hours and is power cycled at the end of the week.
      Which drive lasts longer?
      Obviously drive 1 because the start stop cycles are the same but drive 2 has 7 times as many hours.

      I just know from dealing with industrial motors and servos that starting and stopping is the hardest thing on a motor whether it is big or small due to inrush current. A lot of motors will pull 3 to 4 times their rated current when starting. When I switched our motors over to inverted control (which limits motors to 50% inrush current at start) at my last job, we saw the average life double. We also had many pumps - some that power cycled on a level control and other that ran non-stop. The pumps that power cycled last about 1/3 as long as the pumps that ran non-stop.
      omv 4.0.5 arrakis | 64 bit | 4.12 backports kernel | omvextrasorg 4.0.2
      omv-extras.org plugins source code and issue tracker - github.com/OpenMediaVault-Plugin-Developers

      Please don't PM for support... Too many PMs!
    • Well, for those on the forum, sleeping drives may save a bit of power. But, no matter how one looks at it, sleeping drives doesn't do much to extend drive lives.

      I'm aware of the in rush of current in larger motors, where they actually look like a dead short to the line for a split second. Startup (and the capacitors used in the process for AC motors) is where the highest torque stresses are likely to occur versus idling along at their rated speed. It stands to reason, at least to some degree, that the same would be true of very small motors.

      From what I know of the topic, however, thermal shock is a real enemy of very small mechanical devices. Their tolerances are already very small, so the expansion caused by their own heat and contraction when they cool down magnifies material defects, sets up issues with work hardening, etc. Similarly, thermal shock is one of the major reasons why solid state devices (electronics) fail.
      _________________________________________________________

      My "thesis" on drive longevity falls apart with the properties of the platters, areal density, and "bitrot". Around the 10 year mark, without being refreshed (or rewritten), bits become weaker in magnetic strength and become more subject to read errors. Also, since perfect magnetic media doesn't exist, bits had a greater tendency to "flip" on their own over time. This is why I'm really interested in BTRFS, it's scrubbing processes, and in a potential BTRFS RAID1 pair where bitrot could be repaired.

      ((I've been following the BTRFS project and their RAID1 implementation is still in Beta. Bummer.)) As drives get larger and areal densities increase, bitrot issues with age will become more of a problem.

      With that said, disconnecting power does have the advantage of preventing power mishaps, another major killer of electronics and small devices (surges, dips, spikes, brownouts, etc.). In my case, such an event is less likely because I have whole house surge suppression and independent protection on my servers. Still, having a copy of my stuff disconnected from the wall outlet is insurance against known problems.

      Unless something truly odd and catastrophic happens, I have no plan to restore the data on that very old server. While I don't know what the extent of it is, bitrot must be a factor at this point. (And, with OMV and a new server in the mix, I've recently given thought to pitching it. It's definitely obsolete,) Right now, it's more of an experiment that I know will fail in a few more years. (But,, if BTRFS RAID1 comes out of beta..... :D )

      So, in the final analysis, I agree with you. There are good (power related) reasons to have a cold server but, for the most part, 5 years on average is about the most one could expect of a hard drive. At that point, in most cases, they should be replaced.
      Good backup takes the "drama" out of computing
      ____________________________________
      OMV 3.0.81 Erasmus
      ThinkServer TS140, 12GB ECC /
      32GB USB3.0 / 4TB SG / 1TB WD

      OMV 3.0.81 Erasmus - Rsync'ed Backup Server
      R-PI 2 $29 / 16GB SD Card $8 / Real Time Clock $1.86
      4TB WD My Passport $119
    • flmaxey wrote:

      I'm aware of the in rush of current in larger motors, where they actually look like a dead short to the line for a split second. Startup (and the capacitors used in the process for AC motors) is where the highest torque stresses are likely to occur versus idling along at their rated speed. It stands to reason, at least to some degree, that the same would be true of very small motors.
      When a motor with a very large load tries to start, I've seen the inrush current last for 15 seconds. I'm guessing the same would apply to hard drives with lots of platters vs a drive with one platter.
      omv 4.0.5 arrakis | 64 bit | 4.12 backports kernel | omvextrasorg 4.0.2
      omv-extras.org plugins source code and issue tracker - github.com/OpenMediaVault-Plugin-Developers

      Please don't PM for support... Too many PMs!
    • geaves wrote:

      1. Well I would agree and disagree with both, I've always looked at it from two angles, if it's a server leave it one, if it's a workstation turn it off. In the case of the Pi as this is going to do a specific job i.e. rsync I would probably only run once a week so in that case it would be powered on as and when required.
      2. I am sure the knowledge between the two of you far outweighs my own, but I've always worked on two principles; if ain't broke don't fix it, if it's not needed turn it off.

      I have dealt with two IT companies over the last few years, one advocated shut everything down other than the server, the second suggested that workstations and the server should be left but the monitor switched off.

      3. I would agree that powering a hard drive unnecessarily on and off during a day is going to reduce it's lifespan, but technology has moved forward so much since the early days that pc components are far more robust. I have a 3.5" Maxtor 80Gb which I replaced for a friend, the date on the drive is 2001 but it still works, albeit it sounds like a bag of bolts when it's use.
      1. I have learned one thing regarding the PI and BTRFS, already.
      - Running a scrub on 1TB of data takes hours. (It's a slow CPU.)
      - It's best not to automate a shutdown that might occur in the middle of a scrub (because)
      - I got this error "parent transid verify failed on 14265458688 wanted 464230 found 464221" (which)
      - Forced me to zero out the BTRFS log and do a recovery mount. (This was a bit on the uncomfortable, "first time", side.)
      I think I may dump the automatic AC switch and handle the PI manually. (Replicate once every month or so.)

      2. I don't know about the "knowledge statement", at least where I'm concerned. I was a site admin once, back in the day and on the other side of the pond (but it was 1 of about 5 roles I had at the time). The architecture was an unholy one with an OS2 server, and clients running Win 3.11 for Workgroups. Odd, but it worked fine, especially for that era.
      (That was right around the time when I started testing NT 3.51 and Redhat Linux at home. Ohh,, ahh.. :D ) Like most things, it's a use it or lose it proposition so, in the current day, I'm back to being a computer NOOB. I imagine that you and ryecoarron are more up-to-date, and know a few more things about LAN "stuff" than I do.


      3. BTW: I would have loved to see the SMART stat's for that 80Gb drive. On the other hand, it might not even support SMART as we know it today.
      ____________________________________________________________

      Having turned this over in the mind a few times, thanks to this thread, I think I'm gong to keep the old backup server going just to see how long it lasts. To refresh the "bits on the platters", I think I'll erase the data that's on it now and simply rewrite each folder. If the older unchanged bits are fading, a complete rewrite should restore magnetic strength. (That's my theory.)
      ____________________________________________________________

      So what happened with your R-PI? Is it up, Rsync'ing and, otherwise, doing what you want it to do?
      Good backup takes the "drama" out of computing
      ____________________________________
      OMV 3.0.81 Erasmus
      ThinkServer TS140, 12GB ECC /
      32GB USB3.0 / 4TB SG / 1TB WD

      OMV 3.0.81 Erasmus - Rsync'ed Backup Server
      R-PI 2 $29 / 16GB SD Card $8 / Real Time Clock $1.86
      4TB WD My Passport $119
    • flmaxey wrote:

      The architecture was an unholy one with an OS2 server, and clients running Win 3.11 for Workgroups. Odd, but it worked fine, especially for that era.
      (That was right around the time when I started testing NT 3.51 and Redhat Linux at home. Ohh,, ahh.. ) Like most things, it's a use it or lose it proposition so, in the current day, I'm back to being a computer NOOB. I imagine that you and ryecoarron are more up-to-date, and know a few more things about LAN "stuff" than I do.
      I don't know so much, I remember the days of 3.51 and Win 3.11 for Workgroups, it was the bees knees back then, and what about the protocols, NetBeui. IPX/SPX, TCP/IP was only ever used for an internet connection, dial-up modems, whilst the technology has moved forward in both hardware and implementation the concept has changed very little, it's the understanding these days. There is usually a solution to most things, it's the step by step process that takes the time, it's like users having issues with W10 and OMV SMB, it's not OMV it's M/S's implementation of SMB.
      When it comes to networking and PC's nothing works 'out of the box' you need to have 'some' knowledge and understanding, even with myself if it doesn't work I'll research it until I find a solution. Home networking is a breeze compared to working in an enterprise or school environment, there's so much more that can wrong.

      Saying all that I have neighbour with a young family (I have redeployed W7 for him on 2 of his machines) the hard part has been trying to explain how he can improve his wireless networking and why his internet TV starts buffering at night. They have a son who is 11 years old and will start what we call in the UK senior school, for school use they have to but him a Chromebook but he'll need another PC at home to do homework etc. Anyway I have a laptop I no longer use (4GB Ram, Dual Core) I said he can have that but I'm not going to install M/S on it I'll install Linux Mint.....well that produced a blank expression from the neighbour ^^ so I briefly explained Linux....by this time I'm sure he was comatosed ?( anyway we have left it that they will come over this coming week end and I'll explain and show how he can use Linux just as well as M/S for what he will need.

      ______________________________________________________________________________________________

      Which brings me to something I'm going to have to help him with, at present he gets his kids to use KidGoogle, that seems to be his implementation of keeping kids safe on the net.
      However, I'm going to suggest to him 1 of 3 options;

      1. A router that has content filtering
      2. OpenDNS (if his current router supports it)
      3. A proxy server running on perhaps a Pi

      The problem with the first two is the TOR browser as I think this can circumvent both of them, whereas the third option could not. I'm looking at this from his eldest son who I'm sure will make friends that computer savvy and will know how to get around any obstacles.

      _______________________________________________________________________________________________


      flmaxey wrote:

      Having turned this over in the mind a few times, thanks to this thread, I think I'm gong to keep the old backup server going just to see how long it lasts. To refresh the "bits on the platters", I think I'll erase the data that's on it now and simply rewrite each folder. If the older unchanged bits are fading, a complete rewrite should restore magnetic strength. (That's my theory.)
      Well you have nothing to lose, metaphorically speaking ;) I'm always reluctant to throw away, but I have a greater power that influences my decisions....but in that respect I have noticed it's not a two way thing :D All hail SWMBO =O

      _______________________________________________________________________________________________


      flmaxey wrote:

      So what happened with your R-PI? Is it up, Rsync'ing and, otherwise, doing what you want it to do?
      It's rocking!! It does what is says on the tin, I did the dry run then a live one and it works, or it 'seems' too anyway, I'm happy to run this manually as well as it's in my home office.The only thing I have to do is to resolve the permission side as currently this is running with guest with no problems. So I'm going to go back over your previous post about setting this up and create a user that only has read access to the remote share.

      I also installed the reset permissions plugin only because it gives you an extra tab Shared Folder in Use and gives you path information In Use By, which I think is very useful.

      The post was edited 1 time, last by geaves ().

    • Squid could probably run pretty well on an rpi but the issue is that it takes some GB of storage for even 100mbits downlink internet connection, which I assume a few kids with parents would saturate pretty easily.

      The other thing is that running a proxy like that on a dinky little pi would probably introduce some kind of bottleneck on the network... it'd have to be just about dedicated to that task to be sure!

      Squidsafe would do it and most likely Squid itself would save on a good deal of unnecessary bandwidth usage if there are a number of devices on the network.

      I think it's possible to force devices on the network through that proxy and it's even possible to write rules to exclude the parental computers. Combine that with the pi-hole thingy would be a neat little project, I've been toying with.
      :cursing: Intel NUC Kit DE3815TYKHE My NAS build / ARK
      Atom E3815 1.46GHz4GB RAM • 1TB SSHD • 4GB eMMC
      OMV 3.x | OMV-Extras | Flash Memory | Pi-hole

      RPi/3 №1: LibreElec 8.xRPi/3 №2: SickRage, PiVPN, noip2
    • geaves wrote:

      1. I don't know so much, I remember the days of 3.51 and Win 3.11 for Workgroups, it was the bees knees back then, and what about the protocols, NetBeui. IPX/SPX, TCP/IP was only ever used for an internet connection, dial-up modems, whilst the technology has moved forward in both hardware and implementation the concept has changed very little, it's the understanding these days.
      _______________________________________________________
      2. Saying all that I have neighbour with a young family (I have redeployed W7 for him on 2 of his machines) the hard part has been trying to explain how he can improve his wireless networking and why his internet TV starts buffering at night.
      ______________________________________________________________________________________________

      3. Which brings me to something I'm going to have to help him with, at present he gets his kids to use KidGoogle, that seems to be his implementation of keeping kids safe on the net.
      However, I'm going to suggest to him 1 of 3 options;

      1. A router that has content filtering
      2. OpenDNS (if his current router supports it)
      3. A proxy server running on perhaps a Pi
      1. Back in the day, at least in the beginning of networked PC's, I'd have to say that Novel Netware was a good implementation. As I understood it, some of the admin's were actually forgetting "where" their servers were installed.
      - Since my users were happy with the system I ran, I remember a management request to "help out" a Banyan Vines admin. I asked basic questions and when I found out that he didn't have a viable backup, and hadn't even been backing up for 7 months (around the time the last admin departed), I got the he~~ out of there. (My belief is; "You're only as good as your last backup".) That was a time bomb that I didn't want to be near or associated with.
      - I did teach LAN management in the NT 4.0 days and an installer class for upgrading / migrating from NT4 to Server 2000. In both cases, the focus was very narrow fitting exactly how sites and infrastructure was done by the employer. The worst of it was resolving permissions issues between site admins and the domain level. (There was even a "Forest" level but I didn't get into any of that.) In any case, it was more about M$'s implementation of Active Directory than anything else. The hilarious part of it was; I was an enterprise networking stiff, who had upgraded a networking center's core and a server farm "hot" (no down time) with Cisco cert's and all, and at the end of that assignment, I end up in a LAN Man class.

      2. There could be a few plausible reasons for buffering, like "over subscribed" bandwidth and similar bottlenecks during high use periods.

      3. drinks2go has a point in this one. Note that a PI doesn't have a lot of CPU power to spare. PI's do Rsync acceptably, but that's due to a user not being held up by a PI's slowness. (Rsync jobs can take as long as is needed. That's what makes an R-PI a decent, low cost, low power, backup device.)
      Good backup takes the "drama" out of computing
      ____________________________________
      OMV 3.0.81 Erasmus
      ThinkServer TS140, 12GB ECC /
      32GB USB3.0 / 4TB SG / 1TB WD

      OMV 3.0.81 Erasmus - Rsync'ed Backup Server
      R-PI 2 $29 / 16GB SD Card $8 / Real Time Clock $1.86
      4TB WD My Passport $119
    • drinks2go wrote:

      The other thing is that running a proxy like that on a dinky little pi would probably introduce some kind of bottleneck on the network... it'd have to be just about dedicated to that task to be sure!

      flmaxey wrote:

      Note that a PI doesn't have a lot of CPU power to spare. PI's do Rsync acceptably, but that's due to a user not being held up by a PI's slowness. (Rsync jobs can take as long as is needed. That's what makes an R-PI a decent, low cost, low power, backup device.)
      Thanks for that both of you I thought I might be expecting too much of a Pi, further researching I found this on spiceworks and I think the sophos option if I'm able to find a cheap enough machine + extra nic just might be a good solution. They're a nice family and his intentions are in the right direction, but his understanding and knowledge fall way short, the eldest is 11 and the youngest is just 18 months, so he has some big headaches in front of him.
      I used Smoothwall, this was before it went commercial, used it from the first release, those were the days when we had just a dial up connection, so it was front end for internet connection as well as filtering worked well, but I don't think there has been much development since 2014/15 for the home option which is Smoothwall Express.
    • So, if your R-PI is Rsyncing your shares correctly, are you going to use as a backup data storage device only or are you thinking about using it like I am, as a backup server? (To be activated only if the primary crashes and burns.)

      If you're thinking about using it as a backup server, depending on each shares permissions, you may need a few lines in you Rsycn jobs' extra options.
      Good backup takes the "drama" out of computing
      ____________________________________
      OMV 3.0.81 Erasmus
      ThinkServer TS140, 12GB ECC /
      32GB USB3.0 / 4TB SG / 1TB WD

      OMV 3.0.81 Erasmus - Rsync'ed Backup Server
      R-PI 2 $29 / 16GB SD Card $8 / Real Time Clock $1.86
      4TB WD My Passport $119
    • flmaxey wrote:

      So, if your R-PI is Rsyncing your shares correctly, are you going to use as a backup data storage device only or are you thinking about using it like I am, as a backup server? (To be activated only if the primary crashes and burns.)

      If you're thinking about using it as a backup server, depending on each shares permissions, you may need a few lines in you Rsycn jobs' extra options.
      Predominantly as a backup, so that should I lose more than 1 drive in the raid I have something I can restore, having said that I have been rethinking how I have this setup and been reading and researching. I'm coming to the conclusion that with a small investment and some drive updates, mergerfs and snapraid maybe a better option having recently found this the only option I don't need is Docker. But the article got my interest....added to that the kids bought me a Pi 3B for my birthday which I am currently redeploying Kodi, trying out some skins and the new EmbyCon plugin, which works well except it doesn't support Music only Movies and Tv Shows.
      I'm going to try the original Emby plugin because either my network has suddenly gone slow or it's the way EmbyCon works, not sure which yet....had to remember a number of settings to get Kodi to initiate 5.1 surround sound....but I got the basics working.
    • New

      I still haven't worked out why my particular Emby install has a problem with music media. Again, why I hit play, the song is over (less than a second). When I play a mix, the titles flow by at speed as if songs where 1 second long. I haven't had time to mess around with it but it's definitely strange. (It probably has something to do with the way I'm trying to play files from a remote net location.) In any case.
      ___________________________________________

      EmbyCon? Frankly, I wouldn't even be interested in a media player / manager if it didn't do music. Video is OK but, after a movie is seen a few times, it becomes old news. In my view, TV shows have an even shorter lifespan. (I'll exempt the various Star Trek series' episodes from the general rule. :) ) On the other hand, and with a bit of variety, I'll listen to my favorite music tracks more than a few times.
      __________________________________________


      On the Rsync thing: (This assumes that you'll continue to use the R-PI is a backup "file server".)
      I noticed when I Rsync shares (when new files or changes are copied over) that permissions are reset, at the destination, to the default create mask. root:root.

      I used WinSCP and checked the properties of the source shares, to be certain of the proper permissions. Then I matched those permissions, in the Rsync job, in extra options. This has the effect of restoring the permissions that are on the source, at the destination. (I think the "Prteserve Permissions" setting only applies to jobs that are truely local - not forgein volumes - but I don't know for sure.

      In any case, the following is an example of my work around.
      (And it's not really important if you're just using the R-PI as an emergency copy of your files.)

      Good backup takes the "drama" out of computing
      ____________________________________
      OMV 3.0.81 Erasmus
      ThinkServer TS140, 12GB ECC /
      32GB USB3.0 / 4TB SG / 1TB WD

      OMV 3.0.81 Erasmus - Rsync'ed Backup Server
      R-PI 2 $29 / 16GB SD Card $8 / Real Time Clock $1.86
      4TB WD My Passport $119
    • Users Online 1

      1 Guest