[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

20020824: Disk full again - MCIDAS XCD_START.LOG

>From: Mike Voss <address@hidden>
>Organization: SJSU
>Keywords: 200208211540.g7LFerK22645 McIDAS xcd_run


>My /export/home disk on rossby.met.sjsu.edu filled up again because the
>XCD_START.LOG grew rapidly. I was doing a cat /dev/null to the log to keep
>it small, but perhaps not often enough. Here is the place in the log where
>the errors start showing up:
>snip from XCD_START.LOG
>startxcd.k: m0monexe - :DMSFC                                    new PID= 5062
>DMSFC        Starting: Surface Hourly (SAO) Metar Decoder
>DMSYN: m0syndec - unable to write to syn row= 8
>DMSYN: m0syndec -                  synmd num= 55
>DMSYN: m0syndec - unable to write to syn row= 8
>DMSYN: m0syndec -                  synmd num= 55
>as far as I can tell the last two lines repeat for about 350 MB until the disk
>fills up.

This shows that this problem was in the Synoptic decoder.  For some reason
it was unable to write an entry in its output MD file, MDXX0055.  For the
SYN schema, row 8 is the last row (which corresponds to time) in the file.

Interestingly, none of the systems running XCD that I have access to 
experienced this kind of problem.

>For reference here is the past glimpses:


>So, it looks like DMSFC again, but a different problem.

This one was from DMSYN, not DMSFC.

>I'm inclined to turn
>off DMSFC decoding because I'm not really using it, but I thought you may be
>interested in this in case it reveals a more global issue and not just a quirk
>of my system.

I see that you have turned off XCD decoding, but not running XCD altogether.
You may want to comment out the 'exec "xcd_run MONITOR"' line in ~ldm/ldmd.conf
and then stop and restart your LDM.  This will turn off all XCD processes.

>You are more than welcome to log on to my system and investigate if it
>helps your program development for any reason.

I did logon, and, unfortunately, I have no ready explanation for why the
Synoptic decoder was unable to write to the output MD file.  By the time I
looked this morning, the MD file was gone, so I am left looking at the
log file output which simply continues to give the same error.

I did, however, notice that this particular write problem will send the
Synoptic decoder into an infinite loop.  This is what you were seeing.

>The backup of the XCD_START.LOG is as /data1/XCD_START.BAK if you want to
>look at the whole log file.

Thanks.  I did look at the file, and it was helpful to a point.  I deleted
the file after looking at it since it provided no more data than what
you sent in this email.


Thanks for the access to your machine.  I will have to ponder the infinite
loop condition in DMSYN and come up with a solution.


NOTE: All email exchanges with Unidata User Support are recorded in the Unidata inquiry tracking system and then made publicly available through the web. If you do not want to have your interactions made available in this way, you must let us know in each email you send to us.