cannot save multiple files on 4100

[smarsh]

Hi There,

I recently upgraded the OS to 3.4.805... before that, all was OK (i was on 2.x)

I can FTP without a problem.... but as soon as i try to save multiple files to the Shared folder from Windows XP Pro SP2 (all patched to current).... the 2nd or 3rd file (of however many i have) will get stuck and not complete and times-out.... i get a "errno=6" in the log....

This is the case ONLY when when copying files to the share (mapped or not)...

I had 64MB of memory and thought it mught be running out of some resource? i then put in 256MB..... things are a lot faster, but problem is still there. This is the case if i run individual drives (4 shares) or all together as one large RAID5... Is there some tunning issue? i did not think so, since FTP is quite happy...

I was thinking this could be related to the Microsoft KB885250 issue, but my box is up to date patch wise....

Anyone? i am setting this 4100 up for my son's school with more RAM and 4 120GB drives.... i need a rock-steady solution and feel that the upgrade to 3.4.805 may have been a mistake?

Thanks!

Sam

[blue68f100]

Phonix32 is the 4x00 expert.

First check to see if the starting points are the same. "co de info"

Then check to see if your 4100 has had the MB repair as posted on the sticky, top of threads. If you need help with determing if the repairs have been made let me know. Have ran across at least 1 4100 that was marked as having the repair and did not. In most all cases 240gig and larger models have had the repairs done.

Then I would run SpinRite to check the HD out. I am finding that since mfg no longer check for media defect, the smart tech is doing a lot of repairs on the fly. This throws off all of the timing. I had a brand new drive use 50% of the alotted reserve sectors. This as after the 4500 formated and set the OS up. NO DATA execpt the OS and clean data area, brand new HD.

Are you using JVM?

[smarsh]

i am not using JVM.... odd thing is that the problem ONLY started after doing the OS upgrade and does NOT happen during FTP transfers (goes quickly!)....

I am going to try the "disable smb for NT" option under the windows networking advanced menu...see if that helps....

I'm still on the original 30GB x4 setup... i will put in the new 120GBx4 tonight...


Also, this happens with any of the 4 drives alone or all in RAID5 together....

Thanks.

Sam

[blue68f100]

If your unit has not had the MB mods as posted, you will have nothing but problems with larger drives. This is to remove noise.

Try reolading the OS, I've had a couple that required 2.

Moving from v2 to v3 can generate problems. The starting point on the HD's are calculated differently. Reason it a good idea to reformat. This will not cause a problem unless a HD fails.

[Phoenix32]

I doubt this is an issue with having or not having the Mod, but I would still check and be sure it had the Mod before putting in 4 x 120 GB drives.

Okay, since it sounds like the drives do not have any needed data on them, do this,

Step 1. Go to debug and do;

co de format 10000 /reinit
co de format 10008 /reinit
co de format 10010 /reinit
co de format 10018 /reinit

This will put the partition starting positions in the proper place for OS v3 and ensure they are all in the same place.

Step 2. Perform a FULL FACTORY DEFAULT (4 LED flash) hard reset using the procedure found in the WIKI here;

http://wiki.procooling.com/index.php/Snap_Server_Resets

Now I am talking about a HARD RESET not a reset you do from the web interface. Just follow the procedure for a 4 LED flash (pressing the button 4 times). Do Step 3 right after doing the Hard Reset.

Step 3. Flash the OS again using SNAP Assist. Do not use a different OS version (at least at this point). Do not use an old version of SNAP Assist (like used with OS v2), so make sure it is a version at least as new as the OS you are using. And note, I underlined using SNAP assist, do not do this from the web interface.

Step 4. (optional) Just in case that particular part of the OS was corrupt (I doubt it, but), repeat Step 1.


I bet this cures your problem. If not, there may be a more serious issue going on. If that is the case, I have to ask, are you sure it was working properly before the OS update?

Okay, go do that, then come back and let me know what a genious I am (or idiot if it didn't work).

[smarsh]

Originally Posted by Phoenix32

I doubt this is an issue with having or not having the Mod, but I would still check and be sure it had the Mod before putting in 4 x 120 GB drives.

Okay, since it sounds like the drives do not have any needed data on them, do this,

Step 1. Go to debug and do;

co de format 10000 /reinit
co de format 10008 /reinit
co de format 10010 /reinit
co de format 10018 /reinit

This will put the partition starting positions in the proper place for OS v3 and ensure they are all in the same place.

Step 2. Perform a FULL FACTORY DEFAULT (4 LED flash) hard reset using the procedure found in the WIKI here;

http://wiki.procooling.com/index.php/Snap_Server_Resets

Now I am talking about a HARD RESET not a reset you do from the web interface. Just follow the procedure for a 4 LED flash (pressing the button 4 times). Do Step 3 right after doing the Hard Reset.

Step 3. Flash the OS again using SNAP Assist. Do not use a different OS version (at least at this point). Do not use an old version of SNAP Assist (like used with OS v2), so make sure it is a version at least as new as the OS you are using. And note, I underlined using SNAP assist, do not do this from the web interface.

Step 4. (optional) Just in case that particular part of the OS was corrupt (I doubt it, but), repeat Step 1.


I bet this cures your problem. If not, there may be a more serious issue going on. If that is the case, I have to ask, are you sure it was working properly before the OS update?

Okay, go do that, then come back and let me know what a genious I am (or idiot if it didn't work).

I have done full reset and reformatted a few times to RAID5 and individuals.... just not using the co command, but using the interface for the reformats....

I don't understand how FTP would work poerfectly and not the windows share mount.... It did work fine before (i was benchmarking before all the changes to see if i could gain improvements with new OS, RAM, Drives, etc...).... i'll give it a shot tonight....

I get the errno=6 in the log... and the kind of symptoms smells of communication/security to me (since FTP is flawless and fast)... I will also try the disable "SMB for NT" option i found elsewhere.....

Thanks all for your input.... i will try it all!

Sam

[blue68f100]

While you are at it, disable master browser too. This can cause conflicts with MS tring to do it to. It caused problems on my network, more than 1 master browser, not good.

If you really want to do it clean wipe the drives, then re install. The snap will think it has all new hd and load the os from flash ram. Then you will be required to install the help files.

[smarsh]

yep, it was the "SMB for NT" setting..... i disabled it and voila! it works great. Thanks for the tip on the master browser too....bad! i disabled that to!

Sam

P.S. i rebuilt the RAID5 with the new drives and the RAID automatically started using 128MB of the 256MB.... before this RAID build i had a 64MB chip in there and it was only using 16MB for RAID cache.... i thought i would have to do the manual setting on the RAID cache....????

[blue68f100]

Use the "debug memory" cmd to see if the snap is using all of it. The snap only uses low density chips.

As for cache I think it defaults to 1/2 of available ram. Test data shows it does not provide any speed boost. But if your using JVM it will be a boost.

[Phoenix32]

Originally Posted by smarsh

I have done full reset and reformatted a few times to RAID5 and individuals.... just not using the co command, but using the interface for the reformats....

The co de format 100xx /reinit command does more than the format from the basic admin menu does. When you put in the new drives, it won't matter, they will get formatted all the way.

The reset was just to prepare it for another OS install. Kind of a back to ground zero type deal.

Sorry I was no help, but glad you got it figured out...


Well David, we now know the v3 OS SAMBA has/may have issues with Win2k. That is a new one on me. I have not seen anyone have this problem before with Win2k.

[blue68f100]

He upgraded from v2... Makes me wonder if he is up todate on patches to the OS. Wonder if the 807 had fixes related to his problem.

I'm discovering the more equipment I bring online, more suttle things change, like master browser. I had 4 nas tring to be master. I picked this up on my 2200 log. My APC Network card broadcasts status constantly to hardware on my net devices. The Guardian OS reports this, Snap OS NOT. I picked up this traffic on my personal FW.

[smarsh]

i am on XP Pro not 2K.... did not have access to 807 only found 805...

Thanks all for your help.

Sam

[Lethe]

The 'disable SMB for NT' is what Snap techs' was saying to temporary fix the kb885250 issue, so this does sound a similar problem. MS also issued an update to kb885250 (with a different kb #, just to catch us out) that caused similar issues.

At work I blacked both these patches out until I could flash the Snap boxes up to version 4.x.

But having said that, it was only XP sp1 machines that got the error - sp2 machines worked OK with kb885250 applied though?

Nick

This was the other hotfix that caused (me) problems --> http://www.microsoft.com/technet/sec.../MS06-030.mspx

[smarsh]

yeah, that's where i got the idea from SnapTech..... my XP Pro box is patched up to the latest as of last night..... i do have other network related installs on it like VM and a VPN client that might have affected it.... perhaps... don't know, but disabling it did the trick!

On a side note... i put in 4 new 120GB WD drives and set it up to RAID5 and let it sit over night to test stability..... in the morning i had 1 drive doing the 'clunk of death" that WD is renowned for.... The server did the right thing and continued on it's merry way.... Go figure, a bad WD drive! (sarcasm).

P.S. the 'disable smb for NT' might be a good one for the wiki, as well as disabling the 'master browser' (god knows why they have it on by default).

[Lethe]

Well, turning on 'disable smb for NT' isn't really the way to go - this shouldn't be turned off IMHO.

Have you checked network cables and such - this symptoms you get are typical of a bad network cable etc. I expect disabling SMB for NT reduces network traffic, and thus is a false fix to the real problem?

Disabling SMB for NT shouldn't be in the FAQ either, as it isn't really recommended to do.

Nick

P.S. The actual setting is 'Enable SMB for NT' and on all the 5 Snapboxes at work it is 'ticked'. The help on this specifies to NOT disable it unless explicitly instructed by a Snap Technician.

[smarsh]

the only thing it does, AFAIK, is stop the ability to get over the 4 Gig limit on file size..... other than that, i'm not sure what other affect it has.

[Lethe]

Well, yes, so then why does it 'fix' the issues you have? Something else is awry, I feel, and disabling that feature gets around the issue but really isn't the cause.

Here is a good page about the SMB for NT thing:

http://www.jsifaq.com/SF/Tips/Tip.aspx?id=9344

Nick

[jontz]

Originally Posted by Lethe

Well, turning on 'disable smb for NT' isn't really the way to go - this shouldn't be turned off IMHO.

Have you checked network cables and such - this symptoms you get are typical of a bad network cable etc. I expect disabling SMB for NT reduces network traffic, and thus is a false fix to the real problem?

Disabling SMB for NT shouldn't be in the FAQ either, as it isn't really recommended to do.

If it was simply a bad network cable, I would think that FTP wouldn't work either. Also, it seems from the very first post that this was only a problem writing to the share and not reading from. I would expect to see problems from all of these areas with a bad network cable, not just one particular operation. That being said, it seems like while turning off SMB does work, the fact that it only happens on certain XP machines makes me think the problem does lie with Windows, not the snap server. Imagine that...

[Lethe]

FTP may have a lower overhead than SMB? In fact, I expect it does.

Nick

[jontz]

True, but it seems like there would still be a problem reading and writing, not just one or the other. Hey, stranger things have happened though...

[re3dyb0y]

Originally Posted by jontz

True, but it seems like there would still be a problem reading and writing, not just one or the other. Hey, stranger things have happened though...

Im sure it would be easier to make our own NAS...

[jontz]

Well YEAH, but we aren't interested in the easy approach! That's why we are hobbling 10 + year old hardware along...