|
Home > Archive > SQL Anywhere database replication > August 2005 > ssremote, file protocol and network access
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
| Author |
ssremote, file protocol and network access
|
|
| Gordon.Freeman@Sandtraps.com 2005-07-25, 9:24 am |
| We are using ssremote.exe (as shipped with ASE 12.5.x, set up with FILE
protocol and exchanging at a fixed time interval , continuous execution.
One problem is that the consolidated machine sometimes is re-booted or
down for other reasons. The remote machines use the disk of the
consolidated machine for message file exchange.
I should like to have ssremote report the failing access, but otherwise to
not stop execution. As it is now, the logfile reports error 53 and
execution stops without resuming when the network disk is available again.
I should add that ssremote is wrapped in a Windows Service with the
srvany.exe tool. The service appears to run as normal, while the
underlying ssremote execution has come to a halt.
There is sure a solution to this ?
best regards
Gordon
| |
| Reg Domaratzki \(iAnywhere Solutions\) 2005-07-25, 9:24 am |
| Not being able to write to the messaging system is considered a fatal error
for SQL Remote, and it will retry 5 times ( with a 1,2,3 and 4 second pause
between tries ) before shutting down. If you start ssremote with the -k
switch ( and -o filename.txt ), when "Execution Completed" is printed in the
ssremote log, ssremote will shut down, and the service should detect that
the process has ended.
--
Reg Domaratzki, Sybase iAnywhere Solutions
Sybase Certified Professional - Sybase ASA Developer Version 8
Please reply only to the newsgroup
iAnywhere Developer Community : http://www.ianywhere.com/developer
iAnywhere Documentation : http://www.ianywhere.com/developer/product_manuals
ASA Patches and EBFs : http://downloads.sybase.com/swx/sdmain.stm
-> Choose SQL Anywhere Studio
-> Set "Platform Preview" and "Time Frame" to ALL
<Gordon.Freeman@Sandtraps.com> wrote in message
news:42e4e4c6@forums
-1-dub...
> We are using ssremote.exe (as shipped with ASE 12.5.x, set up with FILE
> protocol and exchanging at a fixed time interval , continuous execution.
>
> One problem is that the consolidated machine sometimes is re-booted or
> down for other reasons. The remote machines use the disk of the
> consolidated machine for message file exchange.
>
> I should like to have ssremote report the failing access, but otherwise to
> not stop execution. As it is now, the logfile reports error 53 and
> execution stops without resuming when the network disk is available again.
>
>
> I should add that ssremote is wrapped in a Windows Service with the
> srvany.exe tool. The service appears to run as normal, while the
> underlying ssremote execution has come to a halt.
>
> There is sure a solution to this ?
>
> best regards
>
> Gordon
>
| |
| Gordon.Freeman@Sandtraps.com 2005-07-26, 3:23 am |
| Thank you for answering me.
We now use the '-k' and '-o' switches . After some testing it turns out
that the service fails if the network is down at service start time. (This
can also be provoked by referencing ssremote.exe over the network). In
this case the service nicely displays with status stopped and we can use
the "Recovery" section of the Windows Service configuration applet to
schedule the behaviour, e.g. re-start after 10 minutes or one hour.
If the network access goes down after some time with succesful execution,
this does not take place. The Service reports status 'Started' while
ssremote terminates completely (does not appear in the task manager). This
may be a weakness in the Windows service wrapper program or in ssremote.
In addition I would like to say that the File Transfer Protocol after all
has to rely on the availability of a network. A network is known to be
unavailable some times, due to re-booting or failures. One would expect
ssremote to offer a recovery scheme for such cases.
regards,
Gordon
| |
| Gordon.Freeman@Sandtraps.com 2005-08-29, 7:23 am |
| OK,
A general problem we have in all cases with SQL Remote is the failure to
detect shut-downs with subsequent filling up of transaction logs. This is
fairly serious for the customers' production environments. It also costs
us a lot of time and money when it happens, in terms of service work.
Assume that the Windows Service wrapper program is not in use at all - I
just want to have ssremote run continuously, without being shut down by a
temporary absense of a message file directory. This is frequently induced
by re-booting on the other side or other types of services on network and
machines. Also note that the repl. frequency is fairly high for my
purpose.
SSRemote shuts down if a write message directory is inaccessible, but not
if a read access directory is inaccessible. When the read directory
becomes available again, ssremote resume reading all right, good.
You say that ssremote tries to re-send 5 times before shutting down
(verified) - perhaps there is a smart way to control the interval between
each try ? That would solve the case. If not, then this is a great
suggestion for next release/version.
In the case with one consolidate and one remote database, a solution is
to place the address given with "sp_remote_type" on the opposite machine.
"sp_grant_remote" and " sp_grant_consolidate
" should then be supplied with
an address on the same machine as where it is issued. Both servers then
write to its own disk and reads from the other machine.
This will however only work in the case of one cons. and one remote. Is
it possible at all to obtain something similar in the general case ?
Of course, I could write a sniffer program that detects the absense of
ssremote and which periodically tries to re-start it. That program in turn
would have to be scheduled and handled for breaks and so on.
best regards
Gordon
| |
| Reg Domaratzki \(iAnywhere Solutions\) 2005-08-29, 9:23 am |
| The best solution for you right now is NOT to run ssremote ( or dbremote )
in continuous mode, but to run in send then close mode and wrap the calls to
ss/dbremote in batch files that loop continuously, possible with a delay
between calls.
--
Reg Domaratzki, Sybase iAnywhere Solutions
Sybase Certified Professional - Sybase ASA Developer Version 8
Please reply only to the newsgroup
iAnywhere Developer Community : http://www.ianywhere.com/developer
iAnywhere Documentation : http://www.ianywhere.com/developer/product_manuals
ASA Patches and EBFs : http://downloads.sybase.com/swx/sdmain.stm
-> Choose SQL Anywhere Studio
-> Set "Platform Preview" and "Time Frame" to ALL
<Gordon.Freeman@Sandtraps.com> wrote in message
news:4312cd4a@forums
-1-dub...
> OK,
>
> A general problem we have in all cases with SQL Remote is the failure to
> detect shut-downs with subsequent filling up of transaction logs. This is
> fairly serious for the customers' production environments. It also costs
> us a lot of time and money when it happens, in terms of service work.
>
> Assume that the Windows Service wrapper program is not in use at all - I
> just want to have ssremote run continuously, without being shut down by a
> temporary absense of a message file directory. This is frequently induced
> by re-booting on the other side or other types of services on network and
> machines. Also note that the repl. frequency is fairly high for my
> purpose.
>
> SSRemote shuts down if a write message directory is inaccessible, but not
> if a read access directory is inaccessible. When the read directory
> becomes available again, ssremote resume reading all right, good.
>
> You say that ssremote tries to re-send 5 times before shutting down
> (verified) - perhaps there is a smart way to control the interval between
> each try ? That would solve the case. If not, then this is a great
> suggestion for next release/version.
>
> In the case with one consolidate and one remote database, a solution is
> to place the address given with "sp_remote_type" on the opposite machine.
> "sp_grant_remote" and " sp_grant_consolidate
" should then be supplied with
> an address on the same machine as where it is issued. Both servers then
> write to its own disk and reads from the other machine.
>
> This will however only work in the case of one cons. and one remote. Is
> it possible at all to obtain something similar in the general case ?
>
> Of course, I could write a sniffer program that detects the absense of
> ssremote and which periodically tries to re-start it. That program in turn
> would have to be scheduled and handled for breaks and so on.
>
>
> best regards
>
> Gordon
>
>
|
|
|
|
|