Pro/Forums - View Single Post

Zavyre · 08-13-2012, 04:53 PM

Hello all,
I’ve been lurking around these forums off and on for a few years but never spoke up. I have learned Quite a bit about some Snap Servers setup and hardware through my own Troubleshooting and from this forum. So I figured it was about time I started imparting what knowledge I can when needed. I have not been posting on many forums much in the last few years, as, well time usually does not permit (hmm … BS … I can make 5-10 of available time in my day to help out.) But in general I do like to figure out the answers to problems myself; makes that moment when you finally get the project working the way you want all that much sweeter

But I’ve noticed a trend in many forums lately (hardware troubleshooting, OCing, & android dev communities) where a few very knowledgeable people are just (understandably) getting burnt out. We can’t lose these people as resources, it would just make life that much harder. So here’s to hoping somebody has a question I can answer

, and maybe help out a bit.

Figured I’d give a bit of a background of my experiences with Snap Servers first
(You can just skip this part if you like, I’m sure nobody really cares, but you never know)

A bit of Background … Back in 2006 I found a very distressed Snap 4100 just sitting on the curb on my way home from work one day. I stopped and approached the owner to see if he was just throwing it out with his trash. He told me his work got rid of it and that it did not work, but I could have it, if I wanted to waste my time trying to fix it. I figured it would be an interesting challenge so I tossed it in my trunk. After getting it home I popped off the top to find a PSU & MB; that’s it! No RAM, No HD’s, No Sleds, not even a CMOS bat. OK this will be interesting. Scrounging through my boxes of miscellaneous compy parts I got a few sticks of PC100/133, a battery, and 2 matching WD120’s. Tested the 120’s in a rig to make sure they were good ... hooked everything up and power it on. It took me a bit to find the NAS on the network and I could not login to it; but it was far from a dead unit. After reading a bit about it on these forums I got the info I needed (TY by the way) to get it up and running. Found a copy of SnapOS 4.0.860 and 2 more WD120’s from a friend, flashed the OS and amazingly the server was up and running in a few days. I took the unit to work and used it up until recently as a personal media/file server for me and my employees. While transfer rates were pretty slow it worked well for streaming music and what not. The server however started acting very weird (stuttering while streaming, and hanging on large file directories.) over the last many months to the point we just stopped using it. After having to talk some with employees and remove GB’s of music and personal files from the 3 workstations I figured I’d better fix that old Snap. The biggest and oddest issue is while you can transfer files to the snap, you will not be able to get them off of it. The file transfer just times out unless you do it 1 (maybe 2) file at a time, and anything over 15MB you can just forget about. Thankfully I did a backup of the server in Jan of 2012 so not all is lost and most of the data is replaceable or just not that important. So I got what I needed/could off the server, restored factory settings, and rebuilt the raid. The server is still acting stupid tried 4 diff DIMMS ranging from 64MB to 256 it was still doing the same thing … looked over the FB1 & FB2 areas on the board again ... it’s a factory repair 003 board. I also tried it as 4 individual drives and still the same thing. So feeling as I defiantly got good use out of a supposedly dead unit that I have just some time and spare parts invested into ... Think it’s time to retire the old girl.

Still needing a solution to my employee data problem I started looking for a new NAS. Not wanting spend very much I ran across a Snap 4500 with 4 250’s in it for $90. Seller claimed it was a working unit but did not accept returns. Knowing that some people would consider it a working unit if it just posted, I was a bit Leary of it. But the photos of it looked like it was brand new. So a supposedly tested working 4500 for $90 plus $30 in shipping ... hmmm. After kicking it around for a few days I figured I would take a chance. It paid off! A like New (I mean no scratches, dust, or discernable signs of ever being hooked up new), in the box with everything but the manual, Snap 4500 shows up a week later. And as expected it posted … and then just hung there indefinitely. It would not recognize the second set of drives or go any further then the loading the SCSI Card BOIS. After letting it sit for at least 30min while I worked on something else I tried it a few more times; noda. I decided to remove the Add-on SCSI Card and booted again. This time going straight into the BIOS (noticing this time that the CMOS Bat had died). Looking over the BIOS I also noticed the only 2 HD’s were showing up, the probability of 2 drives dying in a unit in this condition is Highly improbable so I started worrying that the controller board might be bad. But for giggles I figured I’d test all the drives, so I put a post-it on each drive labeling them 1-4 and started moving the possible bad drives around (yes this will piss off and existing raid but I didn’t care about any of the data on it, if any). All the drives worked as suspected. So wtf. Luckily I had a new CR2032 my desk draw; so I shut the server down and swapped the dead battery out, booted, restored defaults and set the time/date, saved settings and rebooted it again this time without the add-on card. And it worked just fine.

The odd thing about this unit is that it looks like it’s has 2 distinct IDE controllers, as after completing the initial post fining 2 drives it loads another IDE interface (I’m assuming) and then finds the other 2 drives (I have not really looked into the specifics what is actually going on here as I need it running sooner than later, maybe some can tell me). I fortunately or unfortunately forgot to put the drives back in order before I rebooted (oops) but it booted none the less to the “You have to manage the server from the web interface screen.” Interesting and Sweet it works … powering down the server I decided to put the SCSI card back in and see if it was the possible cause of the first couple of boot hangs. After reordered the drives back to their original positions I tried it again and voila it worked. As I did not want to mess with finding the server on an odd address, I finally found the reset button and was good to go. The server had Version 3.X on it which surprised me as my old 4100 had a newer OS on it then that. Any as many of you know finding a solid copy of the 5.2 GSU file is not an easy task, but is ultimately doable. So after the upgraded OS (very simple to do) and a few hiccups with a couple of the drives not reading there usable size at all (probably do to booting with the drives moved around) (FYI for those that do not know – just take the problem drive out and use a good partition manager to wipe all partitions and the remake one big new one, Quick format, and plug back in). All is good and man this thing smokes that 4100 (as well it should). But at ~40+ db it is far from quite, one of my employees has already dubbed it the Hair Dryer (I’m just waiting for him to rename it in windows ... I‘m sure it’s just a matter of time).

Ok so a bit long winded of a read but someone might enjoy it.

Now to the questions.

1. I find it very hard to believe that GuardianOS does not have a simple way of salvaging or undeleting files. I must just be not seeing it as web and forum searches just aren’t pulling up any answers. Is there and extension or something that can be used to recover files that were deleted other then snapshot. Every server I have used has had some nature of a recycle bin up until now. Including the Snap 4100 (iirc been a bit sense i needed to do it). I have not played with SS yet as I really have no need for that nature of a backup on this server. But I do need to be able to undo deleted files, especially in the beginning as we are populating and sorting the new server.

2. Has anyone found any lower db fans that work well for this model server? I am presently running it without the 2 smaller hi-pitched fans and it’s only running marginally hotter well within the specs for a P4. I have also found a few different models of the CPU fan that would probably work just fine at ~ half the noise volume. Also considering using the 2 front fans from the now retired 4100 in place of the 2 aux fans in the 4500. Was mainly curious if anyone has tried to quit one of these units down with success. If not I’ll report back with my results after/went I get around to messing with it.

3. Crude, I can't remember the third oh well I'll edit it later.

I'll try and post a pic later of my fix for lack of HD trays in the 4100. I think it will give some of you a good chuckle, but hey it worked for ~6 years like that.

08-13-2012, 04:53 PM	#1
Zavyre Cooling Neophyte Join Date: Aug 2009 Location: Illinois Posts: 10	Introduction & A Few Questions Hello all, I’ve been lurking around these forums off and on for a few years but never spoke up. I have learned Quite a bit about some Snap Servers setup and hardware through my own Troubleshooting and from this forum. So I figured it was about time I started imparting what knowledge I can when needed. I have not been posting on many forums much in the last few years, as, well time usually does not permit (hmm … BS … I can make 5-10 of available time in my day to help out.) But in general I do like to figure out the answers to problems myself; makes that moment when you finally get the project working the way you want all that much sweeter But I’ve noticed a trend in many forums lately (hardware troubleshooting, OCing, & android dev communities) where a few very knowledgeable people are just (understandably) getting burnt out. We can’t lose these people as resources, it would just make life that much harder. So here’s to hoping somebody has a question I can answer , and maybe help out a bit. Figured I’d give a bit of a background of my experiences with Snap Servers first (You can just skip this part if you like, I’m sure nobody really cares, but you never know) A bit of Background … Back in 2006 I found a very distressed Snap 4100 just sitting on the curb on my way home from work one day. I stopped and approached the owner to see if he was just throwing it out with his trash. He told me his work got rid of it and that it did not work, but I could have it, if I wanted to waste my time trying to fix it. I figured it would be an interesting challenge so I tossed it in my trunk. After getting it home I popped off the top to find a PSU & MB; that’s it! No RAM, No HD’s, No Sleds, not even a CMOS bat. OK this will be interesting. Scrounging through my boxes of miscellaneous compy parts I got a few sticks of PC100/133, a battery, and 2 matching WD120’s. Tested the 120’s in a rig to make sure they were good ... hooked everything up and power it on. It took me a bit to find the NAS on the network and I could not login to it; but it was far from a dead unit. After reading a bit about it on these forums I got the info I needed (TY by the way) to get it up and running. Found a copy of SnapOS 4.0.860 and 2 more WD120’s from a friend, flashed the OS and amazingly the server was up and running in a few days. I took the unit to work and used it up until recently as a personal media/file server for me and my employees. While transfer rates were pretty slow it worked well for streaming music and what not. The server however started acting very weird (stuttering while streaming, and hanging on large file directories.) over the last many months to the point we just stopped using it. After having to talk some with employees and remove GB’s of music and personal files from the 3 workstations I figured I’d better fix that old Snap. The biggest and oddest issue is while you can transfer files to the snap, you will not be able to get them off of it. The file transfer just times out unless you do it 1 (maybe 2) file at a time, and anything over 15MB you can just forget about. Thankfully I did a backup of the server in Jan of 2012 so not all is lost and most of the data is replaceable or just not that important. So I got what I needed/could off the server, restored factory settings, and rebuilt the raid. The server is still acting stupid tried 4 diff DIMMS ranging from 64MB to 256 it was still doing the same thing … looked over the FB1 & FB2 areas on the board again ... it’s a factory repair 003 board. I also tried it as 4 individual drives and still the same thing. So feeling as I defiantly got good use out of a supposedly dead unit that I have just some time and spare parts invested into ... Think it’s time to retire the old girl. Still needing a solution to my employee data problem I started looking for a new NAS. Not wanting spend very much I ran across a Snap 4500 with 4 250’s in it for $90. Seller claimed it was a working unit but did not accept returns. Knowing that some people would consider it a working unit if it just posted, I was a bit Leary of it. But the photos of it looked like it was brand new. So a supposedly tested working 4500 for $90 plus $30 in shipping ... hmmm. After kicking it around for a few days I figured I would take a chance. It paid off! A like New (I mean no scratches, dust, or discernable signs of ever being hooked up new), in the box with everything but the manual, Snap 4500 shows up a week later. And as expected it posted … and then just hung there indefinitely. It would not recognize the second set of drives or go any further then the loading the SCSI Card BOIS. After letting it sit for at least 30min while I worked on something else I tried it a few more times; noda. I decided to remove the Add-on SCSI Card and booted again. This time going straight into the BIOS (noticing this time that the CMOS Bat had died). Looking over the BIOS I also noticed the only 2 HD’s were showing up, the probability of 2 drives dying in a unit in this condition is Highly improbable so I started worrying that the controller board might be bad. But for giggles I figured I’d test all the drives, so I put a post-it on each drive labeling them 1-4 and started moving the possible bad drives around (yes this will piss off and existing raid but I didn’t care about any of the data on it, if any). All the drives worked as suspected. So wtf. Luckily I had a new CR2032 my desk draw; so I shut the server down and swapped the dead battery out, booted, restored defaults and set the time/date, saved settings and rebooted it again this time without the add-on card. And it worked just fine. The odd thing about this unit is that it looks like it’s has 2 distinct IDE controllers, as after completing the initial post fining 2 drives it loads another IDE interface (I’m assuming) and then finds the other 2 drives (I have not really looked into the specifics what is actually going on here as I need it running sooner than later, maybe some can tell me). I fortunately or unfortunately forgot to put the drives back in order before I rebooted (oops) but it booted none the less to the “You have to manage the server from the web interface screen.” Interesting and Sweet it works … powering down the server I decided to put the SCSI card back in and see if it was the possible cause of the first couple of boot hangs. After reordered the drives back to their original positions I tried it again and voila it worked. As I did not want to mess with finding the server on an odd address, I finally found the reset button and was good to go. The server had Version 3.X on it which surprised me as my old 4100 had a newer OS on it then that. Any as many of you know finding a solid copy of the 5.2 GSU file is not an easy task, but is ultimately doable. So after the upgraded OS (very simple to do) and a few hiccups with a couple of the drives not reading there usable size at all (probably do to booting with the drives moved around) (FYI for those that do not know – just take the problem drive out and use a good partition manager to wipe all partitions and the remake one big new one, Quick format, and plug back in). All is good and man this thing smokes that 4100 (as well it should). But at ~40+ db it is far from quite, one of my employees has already dubbed it the Hair Dryer (I’m just waiting for him to rename it in windows ... I‘m sure it’s just a matter of time). Ok so a bit long winded of a read but someone might enjoy it. Now to the questions. 1. I find it very hard to believe that GuardianOS does not have a simple way of salvaging or undeleting files. I must just be not seeing it as web and forum searches just aren’t pulling up any answers. Is there and extension or something that can be used to recover files that were deleted other then snapshot. Every server I have used has had some nature of a recycle bin up until now. Including the Snap 4100 (iirc been a bit sense i needed to do it). I have not played with SS yet as I really have no need for that nature of a backup on this server. But I do need to be able to undo deleted files, especially in the beginning as we are populating and sorting the new server. 2. Has anyone found any lower db fans that work well for this model server? I am presently running it without the 2 smaller hi-pitched fans and it’s only running marginally hotter well within the specs for a P4. I have also found a few different models of the CPU fan that would probably work just fine at ~ half the noise volume. Also considering using the 2 front fans from the now retired 4100 in place of the 2 aux fans in the 4500. Was mainly curious if anyone has tried to quit one of these units down with success. If not I’ll report back with my results after/went I get around to messing with it. 3. Crude, I can't remember the third oh well I'll edit it later. I'll try and post a pic later of my fix for lack of HD trays in the 4100. I think it will give some of you a good chuckle, but hey it worked for ~6 years like that.