Re: invalid message format and I/O error while comunicating with backend - Mailing list pgsql-jdbc
From | Sergi Vera |
---|---|
Subject | Re: invalid message format and I/O error while comunicating with backend |
Date | |
Msg-id | 481EAF91.8020203@emovilia.com Whole thread Raw |
In response to | Re: invalid message format and I/O error while comunicating with backend (Kris Jurka <books@ejurka.com>) |
List | pgsql-jdbc |
Yes, your are right<br /><br /> It is a problem related with multiple threads working on the same Prepared Statement.<br/><br /> I have modified the code to run with one single thread and it runs very smoothly. <br /> Previouslythere were more than one thread working onthe same prepared statements (they were created at server startup) Evenif the code already included some synchronization code, it seems I was loosing control of the threads.<br /><br /> Iwill modify the code, and make sure every thread has his own Prepared Statements.<br /><br /> I thought that sharing thesame prepared statements was more efficient.<br /><br /> Thanks<br /><br /><br /> Kris Jurka escribió: <blockquote cite="mid:481901F4.1070302@ejurka.com"type="cite">This appears to be a thread safety related problem. I believe your codehas one thread setting the parameter values and another thread executing the prepared statement at the same time. Theexecutor does two passes through the parameter list, once to calculate a total message length and another time to sendthe values. If the contents change between the length calculation and the message sending we'll have the wrong lengthand the whole client-server communication gets messed up. The attached test case demonstrates this failure mode. <br/><br /> I'm unsure how hard we should try to fix this, there are a couple of approaches: <br /><br /> 1) Do nothing. It's really a client problem and they shouldn't be setting and executing at the same time. <br /><br /> 2) Justcopy the parameters at execution time so we get a consistent view of them. This may not be exactly what the user wantsthough if the order things actually execute is: execute, set, copy instead of execute, copy, set. <br /><br /> 3) Gothrough all the PreparedStatement functions making most of them synchronized so that you cannot set while an execute isrunning. <br /><br /> Kris Jurka <br /><br /> Sergi Vera wrote: <br /><blockquote type="cite">Hi! <br /><br /> I've beena little busy thoose days and was unable to work on this, but I've made the tcpdump session that you requested and <br/> here are the results <br /><br /><br /> Kris Jurka escribió: <br /><blockquote type="cite">Sergi Vera wrote: <br /><blockquotetype="cite">Thanks Kris for the help <br /><br /> Adding loglevel=2 dind't add any more info on logs, and itwill be not easy to make a self contained program, but I have attached the result of <br /></blockquote><br /> The loglevel=2logging will go to the driver's System.out not into the server error log. <br /><br /><blockquote type="cite">tcpdump-vvv -i lo -w pgsqlerror2.dat <br /><br /></blockquote><br /> This only captures the start of each packetso it doesn't have the whole thing. Could you recapture with: <br /><br /> tcpdump -n -w pgsqlerror3.dat -s 1514 -iany tcp port 5432 <br /><br /> This ups the capture size (-s 1514) and also filters out the unrelated UDP traffic you'vegot going on. <br /><br /><blockquote type="cite">Browsing through the first failing pgsql data chunk, one can seethat: <br /><br /><a class="moz-txt-link-freetext" href="http://img139.imageshack.us/my.php?image=pantallazolm8.png">http://img139.imageshack.us/my.php?image=pantallazolm8.png</a><br /><br/> The last data has column lenght -1 which seems strange even if I don'k know anything of this particular protocol<br /><br /></blockquote><br /> -1 length indicates a NULL value, so that's expected. <br /><br /></blockquote><br/><br /></blockquote><br /><pre wrap=""><pre wrap=""> <hr size="4" width="90%" /> </pre></pre></blockquote><br /><br /><div class="moz-signature">-- <br /></div><a href="http://www.emovilia.com/"> <img border="0"src="cid:part1.05050409.09000402@emovilia.com" /></a><p style="margin-bottom: 0cm;">Sergio Vera<br /> Rosellón,34, 5 Planta<br /> 08029 Barcelona<br /> tel. 902101870<br /><a class="moz-txt-link-abbreviated" href="http://www.emovilia.com">www.emovilia.com</a><br/><br /><p lang="ca-ES" style="margin-bottom: 0cm;"><font face="Consolas,monospace" size="2" style="font-size: 10pt;">IMPRIME ESTE EMAIL Y SUS ARCHIVOS SI REALMENTE LOS NECESITAS.<br/> GRACIAS POR RESPETAR EL MEDIO AMBIENTE.</font><p lang="ca-ES" style="margin-bottom: 0cm;"><font face="Consolas,monospace"><font size="2" style="font-size: 10pt;">NOTA: La información contenida en este email, y sus documentosadjuntos, es confidencial y para uso exclusivo de la persona o personas a las que va dirigido. No está permitidoel acceso a este mensaje a otra persona distinta a los indicados. Si no es uno de los destinatarios o ha recibidoeste mensaje por error, cualquier duplicación, reproducción, distribución, así como cualquier uso de la informacióncontenida o cualquiera otra acción tomada en relación con el mismo, está prohibida y puede ser ilegal.</font></font><plang="ca-ES" style="margin-bottom: 0cm;"><p lang="ca-ES" style="margin-bottom: 0cm;"><p lang="ca-ES"style="margin-bottom: 0cm;"><font face="Consolas, monospace"><font size="2" style="font-size: 10pt;">ADVICE:The information in this email as in the attachments is confidential and private for exclusive use of the targetuser group. Access to this message is disallowed to any other than the addressee. If you are not the addressee or youhave been included by mistake, any duplication, reproduction, distribution as any other action relative to this emailis strictly forbidden and might be illegal.</font></font>
pgsql-jdbc by date: