You're correct; I misunderstood/misread the man page.
- Keith
On 05/ 5/11 09:30 AM, Nirmal Agarwal wrote:
Hi Keith
According to my understanding of timeout option from man page,
timeout is not a total time from start to finish. It just set the
value of --dns-timeout --connect-timeout, and --read-timeout same
as the value of --timeout.Also, timeout states inactive time which
doesn't include the time of download.
<snippet>
--timeout=seconds
Set the network timeout to seconds seconds. This is
equivalent to specifying --dns-timeout,
--connect-timeout, and --read-timeout, all at the same
time.
</snippet>
Please correct me if I am wrong.
Thanks
Nirmal
On 05/05/11 21:04, Keith Mitchell wrote:
Hi Nirmal,
The reason I suggested possibly switching to read-timeout from
an overall timeout is that I'm concerned that 60 seconds
start-to-finish may not be a large enough window for some of the
files being read. The solaris.zlib file, in particular, is quite
a few MB in size - on a slow enough connection, that could
easily take more than 60 seconds total to read. The man page for
wget seems to imply that read-timeout would be more valuable for
this case; it would mean the file could continue downloading
slowly for a longer time; but if there was 60 seconds of
non-activity, wget would try again or fail.
- Keith
On 05/ 5/11 04:28 AM, Nirmal Agarwal wrote:
Hi Keith, Martin
Thanks for reviewing the changes.
I have modified the changes as suggested.With the current
changes wget will try for 10 times waiting for around 10 min
before it finally gives up.
Webrev : http://cr.opensolaris.org/~nirmal27/7013796/
I have done unit test with options waitretry, tries and
timout. In my test case wget exits in 5 min and 10 sec as
expected.
Thanks
Nirmal
Test case output :
------------------------
$ time wget --waitretry=5 --tries=5 --timeout=60
129.158.212.1
21
--2011-05-05 14:50:17-- http://129.158.212.121/
Connecting to 129.158.212.121:80... failed: Connection timed
out.
Retrying.
--2011-05-05 14:51:18-- (try: 2) http://129.158.212.121/
Connecting to 129.158.212.121:80... failed: Connection timed
out.
Retrying.
--2011-05-05 14:52:21-- (try: 3) http://129.158.212.121/
Connecting to 129.158.212.121:80... failed: Connection timed
out.
Retrying.
--2011-05-05 14:53:24-- (try: 4) http://129.158.212.121/
Connecting to 129.158.212.121:80... failed: Connection timed
out.
Retrying.
--2011-05-05 14:54:28-- (try: 5) http://129.158.212.121/
Connecting to 129.158.212.121:80... failed: Connection timed
out.
Giving up.
real 5m10.054s
user 0m0.002s
sys 0m0.005s
---------------------------------------------------------------------------------------------------------------------------------------------------------------
On 05/04/11 23:25, Martin Widjaja wrote:
Keith - The only thing I am worried about just setting
--read-timeout and --connect-timeout is there might be other
timeout that's covered by the overall --timeout that may not
get set and left to the default wget value of - what is it -
900 sec? If that happens to be the case, then the issue will
come back... intermittently.
Other than that, I agree with the rest of the review
comments.
Additionally, I noticed the sleep was just removed
altogether, but I thought there might be advantage in
waiting/sleeping before next retry. Having a static sleep is
probably not desireable though, so there is probably
advantage in also using --waitretry feature of wget? It
sounds like we can reduce the timeout a bit by letting wget
incrementally add sleep seconds between retries.
Martin
On 5/4/2011 9:04 AM, Keith Mitchell wrote:
Hi Nirmal,
The CR says that the total timeout should be around 10
minutes. Since TIMEOUT and TRIES on lines 54-55 are
unchanged, it looks like it will still end up with 900 sec
(15 min) timeouts, and attempting them 20 times, for a
total of 5 hours of waiting if the file can't be
retrieved.
I imagine the value for TIMEOUT should be closer to 60
seconds, and maybe the parameter to wget should be
--read-timeout, not an overall --timeout
(--connect-timeout may be valuable to set as well). TRIES
can be 20 still, though maybe 10 is sufficient.
- Keith
On 05/ 4/11 05:40 AM, Nirmal Agarwal wrote:
Hi all
Please do a code review of the change for:
7013796 AI shouldn't have a 10 hour timeout when
wget'ing files.
Webrev Location :
http://cr.opensolaris.org/~nirmal27/CR7013796/
I have created image using dc and tested it. It works
fine.
Thanks,
Nirmal
--
Sun, an Oracle Company
Nirmal Agarwal
Solaris Install Group
Revenue Product Engineering (RPE), Systems
|Bangalore |
Oracle is committed to developing
practices and products that help protect the
environment
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
--
Sun, an Oracle Company
Nirmal Agarwal
Solaris Install Group
Revenue Product Engineering (RPE), Systems
|Bangalore |
Oracle is committed to developing
practices and products that help protect the environment
_______________________________________________
caiman-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/caiman-discuss
--
Sun, an Oracle Company
Nirmal Agarwal
Solaris Install Group
Revenue Product Engineering (RPE), Systems
|Bangalore |
Oracle is committed to developing practices and
products that help protect the environment
|