Hi Mark/Joe Actually if I wait enough it is able to process the flowfile. When I start nifi, it is able to process immediately, but the second time I have to wait around 1/2 minute. I see similar behaviour on Nifi 1.12.1, I have to wait second time. I don't see the entry log about *Unable to write to container* but I check my disk and it is almost full đ at least in percentage values:
$ docker exec 373913ea4d00 df -h Filesystem Size Used Avail Use% Mounted on overlay 461G 424G 14G 98% / ... Nothing suspicious in logs. So it is a false alarm, the only problem for me is understanding why it takes so long to split an Avro record (result from ExecuteSQLRecord). Juan On Thu, 18 Mar 2021 at 11:43, Mark Payne <[email protected]> wrote: > Juan, > > How full is your disk that holds your content repository? > The issue that Joe mentioned, I donât think, will cause what youâre > seeing. But I am guessing that your content repository is exerting back > pressure. > > Can you check your logs for the message "Unable to write to container"? I > am guessing that you will see a message along the lines of: > > 2021-03-18 10:26:20,988 INFO [Timer-Driven Process Thread-10] > o.a.n.c.repository.FileSystemRepository Unable to write to container default > due to archive file size constraints; waiting for archive cleanup > > This would indicate that your content repository is filling up and is > exerting backpressure on the Input Port, which prevents it from writing the > data out. This can potentially result in the transfer timing out. > > By default, the content repository exerts backpressure if the disk is 50% > full. This can be configured in nifi.properties: > > nifi.content.repository.archive.max.usage.percentage=50% > > > Thanks > -Mark > > > On Mar 17, 2021, at 11:58 PM, Joe Witt <[email protected]> wrote: > > Thanks Juan - that would be very valuable actually. I'll send you a link > to a build here in an hour or so. If you can test that and let us know that > will help us with the release candidate voting process quite a bit. > > Thanks > > On Wed, Mar 17, 2021 at 8:49 PM Juan Pablo Gardella < > [email protected]> wrote: > >> Wow that is fast! You are awesome, thanks Joe. I will test it. >> >> Juan >> >> On Thu, 18 Mar 2021 at 00:47, Joe Witt <[email protected]> wrote: >> >>> Juan >>> >>> We found a bug in 1.13.1 today as reported here >>> https://issues.apache.org/jira/browse/NIFI-8337 and >>> https://issues.apache.org/jira/browse/NIFI-8334. >>> >>> We will have a 1.13.2 out asap to fix this and the regression now has >>> tests to prevent it in the future. >>> >>> Thanks >>> Joe >>> >>> On Wed, Mar 17, 2021 at 8:44 PM Juan Pablo Gardella < >>> [email protected]> wrote: >>> >>>> Hi all, >>>> >>>> I am using latest nifi version and SplitRecord works only once and then >>>> hangs: >>>> >>>> <image.png> >>>> >>>> I cannot stop it also. >>>> >>>> Juan >>>> >>> >
