Problems with write latency

Hi to all storage gurus,
One of our disk systems (HP EVA8000) has been quite busy lately. HP has
spent days analyzing the situation and the result of the analysis was
that the disk system is ok but there are occasionally up to 2000 ms
write latencies. According to them this is not out of ordinary and
operating systems should be able to handle the situation.
The problem is that Netware seems not always to be able to handle such
write latencies and the clustered services may go to comatose when the
write latency is so high. This does not happen on every such occasion,
but too often anyways. Netware is the only operating system which has
any problems with our EVA8000. In this particular EVA we we have roughly
200 hosts, out of which about 10% are Netware. Thus is seems rather
clear that something is wrong with our Netware setup or Netware itself,
we just dont have any clues what.
All servers have NW6.5 SP8, and HP supported Qlogic HBA's with latest HP
supported bios and drivers.
In our Startup.ncf we have:
set multi-path support=on
LOAD SCSIHD.CDM AEN TOFACTOR=5
LOAD QL2X00.HAM SLOT=201 /LUNS /ALLPATHS /PORTNAMES /MAXLUNS=12
/XRETRY=400 /XTIMEOUT=60 /BACKOFF
LOAD QL2X00.HAM SLOT=202 /LUNS /ALLPATHS /PORTNAMES /MAXLUNS=12
/XRETRY=400 /XTIMEOUT=60 /BACKOFF
The error messages are as follows:
18.06.2009 15.30.14 : COMN-3.27-34
Severity = 5 Locus = 2 Class = 6
NSS-2.70-5005: Volume KOTI12/USER2 user data write
(20204(zio.c[2301])) to block 122610663(file block 4810)(ZID
146302323) failed.
18.06.2009 15.30.15 : COMN-3.27-1092
Severity = 4 Locus = 3 Class = 0
NSS-3.00-5001: Pool KOTI12/SARKA is being deactivated.
An I/O error (20204(zio.c[2279])) at block 29505990(file block
104772)(ZID 3) has compromised pool integrity.
Any ideas what could be wrong?
Best regards,
Jonne Viljanen
University of Helsinki

Hi JJ,
Thanks for your insightful answer. We are using execution throttle 16 on
almost all hosts. However HP suggested to increase the execution
throttle on Netware hosts to 64, since their IO to disk system is very
small and thus increasing the value would not cause problems to other
hosts. Increasing the execution throttle did not solve our problems.
What value are you using for execution throttle?
Cheers,
JV
> Hi,
>
> We had some similar problems (many hosts on an EVA 8000) but only with a
> disk failure. A disk failure causes latency and thus resulting in node cast
> out, comatose services or dismounted volumes. And since a san is still is a
> shared disk system /EXECTHROTTLE might be something to look into. If a
> couple of hosts have the execution throttle set to high (like 256) then they
> will fil up the controllers queues first leaving no room for other hosts.
> We also we played around with the /XTIMEOUT parameter. but not as high as 60
> (ours is 8 on a 8 node cluster) because the cluster software also needs some
> time to make a good judgement.
>
> Cheers,
> JJ
> .
>
>
>
>
> "Jonne Viljanen" <jonne.viljanen@rmv_this.helsinki.fi> wrote in message
> news:[email protected]...
>> Hi to all storage gurus,
>>
>>
>> One of our disk systems (HP EVA8000) has been quite busy lately. HP has
>> spent days analyzing the situation and the result of the analysis was that
>> the disk system is ok but there are occasionally up to 2000 ms write
>> latencies. According to them this is not out of ordinary and operating
>> systems should be able to handle the situation.
>>
>> The problem is that Netware seems not always to be able to handle such
>> write latencies and the clustered services may go to comatose when the
>> write latency is so high. This does not happen on every such occasion, but
>> too often anyways. Netware is the only operating system which has any
>> problems with our EVA8000. In this particular EVA we we have roughly 200
>> hosts, out of which about 10% are Netware. Thus is seems rather clear that
>> something is wrong with our Netware setup or Netware itself, we just dont
>> have any clues what.
>>
>> All servers have NW6.5 SP8, and HP supported Qlogic HBA's with latest HP
>> supported bios and drivers.
>>
>> In our Startup.ncf we have:
>>
>> set multi-path support=on
>> LOAD SCSIHD.CDM AEN TOFACTOR=5
>> LOAD QL2X00.HAM SLOT=201 /LUNS /ALLPATHS /PORTNAMES /MAXLUNS=12
>> /XRETRY=400 /XTIMEOUT=60 /BACKOFF
>> LOAD QL2X00.HAM SLOT=202 /LUNS /ALLPATHS /PORTNAMES /MAXLUNS=12
>> /XRETRY=400 /XTIMEOUT=60 /BACKOFF
>>
>>
>> The error messages are as follows:
>>
>> 18.06.2009 15.30.14 : COMN-3.27-34
>> Severity = 5 Locus = 2 Class = 6
>> NSS-2.70-5005: Volume KOTI12/USER2 user data write
>> (20204(zio.c[2301])) to block 122610663(file block 4810)(ZID 146302323)
>> failed.
>>
>> 18.06.2009 15.30.15 : COMN-3.27-1092
>> Severity = 4 Locus = 3 Class = 0
>> NSS-3.00-5001: Pool KOTI12/SARKA is being deactivated.
>> An I/O error (20204(zio.c[2279])) at block 29505990(file block
>> 104772)(ZID 3) has compromised pool integrity.
>>
>>
>> Any ideas what could be wrong?
>>
>>
>> Best regards,
>>
>> Jonne Viljanen
>> University of Helsinki
>
>

Similar Messages

Data format problem with Write to Spreedsheet File

I have a problem with the data format with Write to spredsheet file.
I have an N*3 array, where the 3 elements in each row should has different length. When I used Write to spreedsheet file,
it can only save the three data in one format. How can I save them in the format of "%0.8f %0.3f %0.2f"?
Thanks for your help.

Hi powerplay,
another solution may be to convert column-wise using "number to fractional string", then interleaving resulting arrays and again using "array to spreadsheet string" (with "space" as separator).
Many ways lead to Rome
Best regards,
GerdW
CLAD, using 2009SP1 + LV2011SP1 + LV2014SP1 on WinXP+Win7+cRIO
Kudos are welcome

Problems with write to measurement

My project group has some problems with the write to measurement VI.
Firstly, for each of the SOC/Voltage...each file only has one value, how to make all value inside one file?
Secondly,How to make all the different parameters, such as SOC1, SOC2, Voltage1, Voltage2, appear in one excel file for easier reference?
Attachments:
TwoBatteries.vi ‏246 KB

Look at your settings for the Write to Measurement File. You are telling it to save to one file, and then use next available filename. Try append to file.
Also, use Merge Signals to combine the 4 and write to a single Write to Measurement File. Right now you have 4 identical ones with the same file name. They are stepping all over each other.
When using the Write to Measurment File, try opening up the help file and reading it o understand what each setting does.

Problems with write to measurement file

My VI doesn't seem to write my signal properly when writing is enables.. however, if i change the "Write to measurement file" Express VI to a Write to Spreadsheet function, the signal is written correctly. From what I understand in previous discussions, there is a bug associated with enabling a write to measurement file, so is there a way to enable write to spreadsheet instead of using the express VI or is there a way I can bypass this bug?
Thanks,
TFlax
Attachments:
Analog test and collection.vi ‏700 KB

This is my second program ever written, first in matlab, so I am very new to the field. So why I choose one way over another I have no reasoning, all I know is that it works. I am collecting electromyography analog signals, with a range of +/-5V.. the "bad" writing with write to meas file is that I am getting a flat signal (range between -.01 to +0.01, give or take).. where when written with write to spreadsheet, the signal has proper voltage range and normal frquency and amplitude characteristics seen in EMGs.
I tried using a case structure for the write to spreadsheet, however it only writes the buffer size during continuous sampling. I need it to write the entire time the boolean (enable button) is true, also the enable button is operated as switch when pressed.
The discussion about the write to measurement file: http://forums.ni.com/ni/board/message?board.id=170&message.id=454380
Thanks again

Problem with Write To Measurement File Express VI not writing to file if file was deleted after the first call (7.1)

I'm using the Express Write to LabVIEW Measurement File in LV7.1.
I have it set to "Save to one file" and have tried both "Overwrite file" and "Using next available file name" for the "If a file already exists" conditions.
The path/filename is passed into the function.
It works fine the first time around. If the file is deleted before the function is called again, an error occurs since LV is unable to find the file.
I have had to stop and re-run my program in order to write to use the function again.
Since I am able to delete the file I assume it was closed but it seems like LV expects it to be there, even though the function was supposed to create/open/close the file... Is there any way to reset this without stopping/restarting the program?
Since my application creates several files (only one using the Express VI) I am checking for the filenames and requiring that they be deleted (or a new filename be used) before acquiring data. Everything has worked great, except for the Express VI.
Thank you,
David

Ok, I've edited the VI you posted to reproduce my error (this time with 8.0).
I added a "master stop" and "start log" button. To demonstrate my problem:
1. Run the VI
2. Click "START LOGGING" - c:\test.txt is created
3. Click "STOP" - c:\test.txt is closed
4. Delete the file c:\test.txt with Windows shell or explorer
5. Click "START LGGING" - a new c:\test.txt file should be created but instead an error is raised in the Write LabVIEW Measurement File function.
Any ideas?
Thanks,
Dave
Attachments:
Write LV Meas File 2.vi ‏96 KB

Problem with Writer.flush() appending extra character

I have written a simple file copy that does the following:
     public static void copy(File source, File dest) throws IOException
          FileReader fr = new FileReader(source);
          FileWriter fw = new FileWriter(dest);
          try
               char[] ch = new char[(int)source.length()];
               fr.read(ch);
               fw.write(ch);
          finally
               fr.close();
               fw.close();
It works fine, and has done for some time.
For some reason, in a recent data collection, I have 3 files out of about a hundred, where the following behaviour occurs.
The char array is written to the file fine. For some reason when fw.flush() is called (i have tested this and confirmed that it is the call to flush() that is causing it) it appends a char to the end of the file with a value of -1.
This is causing merry havoc with my XML parser.
Can anybody tell me how to stop Writer from adding a -1 to my file?
This is on java 1.4.2 currently.

               char[] ch = new char[(int)source.length()];File.length() returns the number of bytes in the file.
However, you're loading that file via a Reader, which converts those bytes into characters using some encoding method. Depending on the encoding, the number of characters produced may be completely different from the number of bytes read. It's also possible that the bytes read are not valid with the encoding.
If you're just copying files, use FileInputStream and FileOutputStream. Better, download the Jakarta Commons IO library, and use their already-tested methods.
If you're working with a JAXP XML parser (ie, what comes with JDK 1.4 and above), create your InputSource from a FileInputStream. That will let the XML parser use the encoding specified in the file, ignoring whatever your JRE's default encoding is.
If you're planning to work with the raw text, then you're going to have to learn about character encodings. I don't have a tutorial link handy, so either Google for one or start with the documentation for java.nio.charset.Charset.

Problem with write a table to excel

I want to open an existing Excel-file from Labview with ActiveX.
I have used the Example "write table XL.vi"
I changed the code that I can open the existing file but I am unable to
select a different sheet in Excel.There is always the first sheet active
and I cannot change this.
Can anybody help me?
robert
Sent via Deja.com http://www.deja.com/
Before you buy.

wrote in message
news:8q4e0a$hsa$[email protected]..
> I want to open an existing Excel-file from Labview with ActiveX.
> I have used the Example "write table XL.vi"
> I changed the code that I can open the existing file but I am unable to
> select a different sheet in Excel.There is always the first sheet active
> and I cannot change this.
> Can anybody help me?
The workbook object has a "Sheets" property that returns a collection of all
the sheets in the active workbook. Use the "Item" method on this collection
to select the sheet you want- if necessary by cycling through all the
elements (use the "Count" property to find how many objects there are) and
using the "Name" property to get the name of the worksheet for comparison
with the one
you're looking for. Remember that everything is case sensitive.

Problem with write key VIs (config-file)

Hi
My problem is this: when I use the write key VIs to write to an ini-file, the values are not written/updated in the ini-file. Does
anybody has a solution???
thank YOU!!!
Attachments:
writetest.vi ‏32 KB

astroboy wrote:
Hi Dennis, writing key updates or add lines, but the problem is that it also causes all lines to be clumped together, making it unreadable.
What's the better way?
Clement
I was complaining about the poor formatting of the .ini files myself a couple years ago in this thread. http://forums.ni.com/ni/board/message?board.id=170&message.id=273552
I found the offending piece of code was buried pretty deep in the Config File VI hierarchy.
One idea is to go back after the file is written and find where all the section headers are and insert your own blank lines.
Also look at this thread http://forums.ni.com/ni/board/message?board.id=170&view=by_date_ascending&message.id=408776
You may also want to add kudoes to the config file idea in the Idea board. http://forums.ni.com/t5/LabVIEW-Idea-Exchange/Comments-in-configuration-file-ini/idi-p/927874#A614
Message Edited by Ravens Fan on 09-17-2009 10:29 PM

Problem with Write Statement

Hi Friends,
Select * from ekko into table i_ekko up to 100 rows.
loop at i_ekko into w_ekko.
write : w_ekko.
endloop.
when i write the above code i am getting an error "cannot be converted in to charater type".
but i want to display the all the coloums and rows from the DB table in the report.
regards
kumar m

hi,
loop at i_ekko into w_ekko
Here w_ekko is considered as the field string which cannot be written with name with reference of the field name.
You shold try this
loop at i_ekko into w_ekko.
write : w_ekko-filed1.
endloop
This will work.
Regards
Sumit Agarwal

TS4094 Have a problem with drums latency when i rec it al seems good but when i playback th drums have a big latency issue please help

Drums latency help

Sounds like the Graphics chip has a problem, which I think is a Mainboard replacement.
You can get a quote for the repair at an Authorized Service Provider. There is an ASP list on the Toshiba website.

Output problem with write() function.

Hi there
This program produced the output only when i add "System.out.write('\n')",
WHY?
import java.io.*;
class WriteBytes
public static void main(String[] args) throws IOException
byte[] b = {'T','e','s','t'};
for(int i = 0; i < b.length; i++)
System.out.write(b);
System.out.write('\n'); // Why if i comment this line then, don't get the output.

Your program is finishing before the buffer has a chance to flush. The newline character is forcing a flush to the print output. You could use System.out.println() to force the newline or System.out.flush() would work as well.

Problem With Write-On Effect

Hi. Help.
I want to use the Write-On effect to animation some script text.
I use the Pen Tool to draw the mask. I Copy the Mask Path, apply the Write On Effect, then I Paste the Mask Path to the Write On's Brusk Position.
The Brush Position keeps winding up to the left of the original path, and a little upwards. The Brush Position is off the Comp, and it seems to animate correctl, just in the wrong place.
I have tried this more than once. What am I doing wrong?
(This is CS3, by the way.)

There are several ways.
One, first paste the mask in a solid which is comp size, and free transform the mask (not the transform properties for tha layer) until it has the same size and location as the original, then copy the mask path and paste into Write-on's Brush Position. Of course, make sure the solid you're using for Write-On is comp size (Set Write-On's Paint Syle menu as "On transparent", so that it composites over layers below).

Latency problem with X-Fi Xtreme Au

Hi all. I'm new in this forum. I have a strange problem. I've bought?a X-Fi Xtreme Audio?soundcard,?and I've noticed a serious latency problem with home recording ap
ps.
With my onboard soundcard I was able to play/record with no latency, using the Asio4all driver. With my new s.card I've a very high latency while playing in realtime. Is it a known problem? Is there any solution? My pc is very?powerful, so I exclude any pc's hardware cause. Thanks in advance, bye

Hi Cardan,
I've got the same Problem with high latency as you with my PC express soundcard with VISTA.
Have you got any solutions?
regards
Mat?

Problems With iTunes Library in Garageband 10.0.1

Hi, If you GarageBand (latest version) fades media and downloads a song from the iTunes Libary in GarageBand, there are problems with the latency! Continuous dropouts and stuttering. But if one loads the same track in GarageBand from the desktop, there are no problems. The loops and effects seem to work when inserting them directly ind GarageBand.

I was finally able to resolve this problem by reinstalling iLife '11. Immediately after installing I had to do a software update to bring iPhoto up-to-date. Can't guarantee this will work for everyone, but it worked for me.
Peter

Audio latency problem with VST, SONAR and Audigy 4 PRO

Hello all,
I have a slight problem with audio performance on my computer and I can't isolate what is wrong. I am using Cakewalk SONAR 4 studio edition, and a recently purchased Audigy 4 Pro soundcard. ( I know it's not quite professional grade, but it works well for me) I recently bought some string instruments from the Vienna Symphonic Library. So far, I love the sound of these "virtual instruments" if you will. The problem comes when I try to record them. I get stuttering and dropouts when I try to record the track in Sonar. I've tried the WDM drivers in 16 and 24 bit modes and the ASIO drivers (which apparently only support 16 bit!?) I know Creative has mentioned that these cards can record easily at 24/96 which is what I'd like if possible. Vienna instruments come in a 24 bit format. Is there something I'm doing wrong or do I need to seriously consider a different soundcard? I use Sonar to bring in the Vienna instruments as a VST. I also apply a reverb on it. When I just play this stuff on my MIDI keyboard, it all sounds great and keeps up. The moment I go to record though, I get dropouts after about 15 seconds or so. Also, is there no way to get ASIO 24 bit support for creative's consumer level cards? I'd be happy with the WDM's personally if those would work. I've looked into the KX project as well but haven't had a chance to try their drivers yet. On a different note, what are the significant differences between ASIO and WDM if any at all? (aside from supported bit rates) My current computer setup is quite powerful and I don't use much CPU power at all so I don't think it's my setup. Intel E6700 (2.67Ghz) Core 2 Duo, 2GB DDR2-800, 320GB Seagate SATA in RAID 1, Nvidia 7600GT, and Creative Audigy 4 Pro. Any help, or insight would be greatly appreciated!
Matthew

Audigy 4 Pro should support 24-bit/96 kHz w/ ASIO --> have you selected those I/O drivers in Sonar audio settings (Options-Audio-Drivers -tab). ! You need to unselect all I/O drivers first to get those 24/96 drivers selectable)?
Dropouts comes from various sources. Have you checked if the read/write caching is set properly (Sonar, try different settings)?
Stuttering may come from shared IRQ/resources and PCI Latency settings.
If your graphics card is not PCIe then, make this registry tweak
Run regedit and search the key named as DisableProgPCILatency
- set all 'DisableProgPCILatency' key values to 1 (HEX or DEC) (1 = disabled, 0 = enabled)
- boot to BIOS settings and find the "PCI latency timer" option --> set its value to 32 or 64 (this is global PCI latency setting)
- reboot
and check the IRQ situation. If there are IRQ shares between Audigy and other devices/controllers, try to get those fixed by disabling unneeded controllers/devices or by moving the Audigy into another PCI slot (PCI slots #3 and #4 are normally not shared).
Couple meaninful differences between WDM/KS and ASIO are;
- you can't get latency lower than 10 ms w/ WDM/KS (well, you can in Sonar by manually editing those buffer values but, the result is not good when you go under 10 ms buffer values (buffersize / samplerate * 1000 = latency in ms --> 512/96000*1000=5.33ms)
and
- you can record from one stereo (you can split this to two mono) source at a time w/ WDM (recording source is set in CT Surround Mixer (or windows recording controls))
- ASIO supports (Audigy serie) locked 'resolutions' only - 16/48 or 24/96
Also, tweak your system for music - http://www.musicxp.net/
Some information on CT ASIO -
http://forums.creative.com/creativel...essage.id=1726
jutapa
Message Edited by jutapa on 12-17-200606:01 PM

Problems with write latency

Similar Messages

Maybe you are looking for