Behavior difference for walsender and walreceiver for n/w breakdown case

Поиск
Список
Период
Сортировка
От Amit Kapila
Тема Behavior difference for walsender and walreceiver for n/w breakdown case
Дата
Msg-id 005d01cd8bf6$206ea430$614bec90$@kapila@huawei.com
обсуждение исходный текст
Список pgsql-hackers
<div class="WordSection1"><p class="MsoNormal"><span style="font-size:10.0pt;font-family:"Arial","sans-serif"">I have
observedthat currently incase there is a network break between master and standby, walsender process gets terminated
immediately,however</span><br /><span style="font-size:10.0pt;font-family:"Arial","sans-serif"">walreceiver detects the
breakageafter long time. </span><br /><span style="font-size:10.0pt;font-family:"Arial","sans-serif"">The main reason I
couldsee is due to replication_timeout configuration parameter, walsender checks for replication_timeout, if there is
nocommunication from other side till replication_timeout time it detects it as a condition to terminate the walsender.
</span><br/><span style="font-size:10.0pt;font-family:"Arial","sans-serif"">However there is no such mechanism in
walreceiver,it fails during send socket call from XLogWalRcvSendReply() after calling the same many times as internally
mightbe in send until the sockets internal buffer is full, it keeps accumulating even if other side recv has not
receivedthe data.</span><br /><br /><span style="font-size:10.0pt;font-family:"Arial","sans-serif"">Shouldn't in
walreceiver,there be a mechanism so that it can detect n/w failure sooner</span>?<br /><br /><br /><span
style="font-size:10.0pt;font-family:"Arial","sans-serif"">BasicSteps to observe above behavior</span><br /><span
style="font-size:10.0pt;font-family:"Arial","sans-serif"">1.Both master and standby machine are connected normally, <br
/>2.then you use the command: ifconfig ip down; make the network card of master and standby down, <br />Observation <br
/>mastercan detect connect abnormal, but the standby can't detect connect abnormal and show a connected channel long
time.<br /><br />With Regards,</span><br /><span style="font-size:10.0pt;font-family:"Arial","sans-serif"">Amit
Kapila</span></div>

В списке pgsql-hackers по дате отправления:

Предыдущее
От: Daniel Farina
Дата:
Сообщение: Re: Proof of concept: standalone backend with full FE/BE protocol
Следующее
От: Amit kapila
Дата:
Сообщение: Re: [WIP PATCH] for Performance Improvement in Buffer Management