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

[IDDBrasil #XSZ-617864]: problems with solon



Hi Guilherme,

re:
> after another power failure last saturday at NCE - UFRJ's datacenter -, solon
> went offline and did not return. Apparently the HDs are taking a long time to
> be checked (if they're being checked at all), something that I'll be
> monitoring closer...

Hmm... this does not sound good.

> After its (forced) return today, the following message appeared on it's log
> several times:
> 
> Apr 03 16:27:14 solon pqact[595] WARN: write(7,,512) to pipe took 2.127325 s
> 
> Is this warning related with fragmented/problematic hard drive, or is it due
> to a network delay?

This kind of warning in the ~ldm/etc/ldmd.log file indicates that it is taking
a long time to process data out of the LDM queue.  This can occur on heavily
loaded systems where there is alot of I/O, or on systems where there is some
problem writing to disk.  It sounds like you might have the latter problem:
slow writes to disk.  What does top, vmstat, or iostat say about disk I/O?

> Also the folks at Brazil's Navy (DHN) complained that they're not receiving
> data from us. The connection was refused by LDM, but there's an allow entry
> for the entire network at DHN:
> 
> allow   ANY     ^200\.244\.76\.*$

What does the log file say about connections being refused?

> I've added individual lines for each host, but it should be working for the
> entire scope, right?

Correct.  The possible reasons for the connection being refused are:

- allows do not include the machine making the request
- datastream being requested is not allowed for the downstream (a refusal for
  one datastream does NOT imply refusal for all datastreams, however)
- a firewall is preventing the request by the downstream from reaching the
  upstream

> By the way, I'll be presenting a paper about the IDD-Brazil at "The 8th
> International Conference on Southern Hemisphere Meteorology and Oceanography"
> this month (April 28th), as I've said before. I'm sending the extended
> abstract to you - I wanted to send it before the submission deadline, but I've
> finished it in the last moment (again :) ). The general idea is to introduce
> the IDD network and the Unidata softwares to the attendees - since it's an
> oral presentation it will be much more interesting :-)

Excellent!

> As a side note, I was very pleased to see that USP, UFPA, CPTEC and
> Universidad de Buenos Aires have accepted abstracts for the AGU Assembly. It
> should be a great opportunity to share some ideas and set the IDD-Brazil in
> perspective (_if_ everyone manages to go!).

I agree completely.  I will be sending you a separate (good) email about your
request for travel assistance.


Cheers,

Tom
****************************************************************************
Unidata User Support                                    UCAR Unidata Program
(303) 497-8642                                                 P.O. Box 3000
address@hidden                                   Boulder, CO 80307
----------------------------------------------------------------------------
Unidata HomePage                       http://www.unidata.ucar.edu
****************************************************************************


Ticket Details
===================
Ticket ID: XSZ-617864
Department: Support IDD Brasil
Priority: Normal
Status: Closed