Обсуждение: BF assertion failure on mandrill in walsender, v13
Hi, Not sure if there is much chance of debugging this one-off failure in without a backtrace (long shot: any chance there's still a core file?), but for the record: mandrill choked on a null pointer passed to GetMemoryChunkContext() inside a walsender running logical replication. Possibly via pfree(NULL), but there are other paths. That's an animal running with force_parallel_mode and RANDOMIZE_ALLOCATED_MEMORY, on AIX with IBM compiler in 32 bit mode, so unusual in several ways. https://buildfarm.postgresql.org/cgi-bin/show_log.pl?nm=mandrill&dt=2021-06-06%2015:37:23
On Thu, Jun 10, 2021 at 10:47:20AM +1200, Thomas Munro wrote: > Not sure if there is much chance of debugging this one-off failure in > without a backtrace (long shot: any chance there's still a core > file?) No; it was probably in a directory deleted for each run. One would need to add dbx support to the buildfarm client, or perhaps add support for saving build directories when there's a core, so I can operate dbx manually.
On 6/10/21 1:47 AM, Noah Misch wrote: > On Thu, Jun 10, 2021 at 10:47:20AM +1200, Thomas Munro wrote: >> Not sure if there is much chance of debugging this one-off failure in >> without a backtrace (long shot: any chance there's still a core >> file?) > No; it was probably in a directory deleted for each run. One would need to > add dbx support to the buildfarm client, or perhaps add support for saving > build directories when there's a core, so I can operate dbx manually. > > This is what the setting "keep_error_builds" does. In the END handler it renames the build and install directories with a timestamp. Cleanup is left to the user. I don't have much knowledge of dbx, but I would take a patch for support. cheers andrew -- Andrew Dunstan EDB: https://www.enterprisedb.com
On Thu, Jun 10, 2021 at 09:08:06AM -0400, Andrew Dunstan wrote: > On 6/10/21 1:47 AM, Noah Misch wrote: > > On Thu, Jun 10, 2021 at 10:47:20AM +1200, Thomas Munro wrote: > >> Not sure if there is much chance of debugging this one-off failure in > >> without a backtrace (long shot: any chance there's still a core > >> file?) > > No; it was probably in a directory deleted for each run. One would need to > > add dbx support to the buildfarm client, or perhaps add support for saving > > build directories when there's a core, so I can operate dbx manually. > > This is what the setting "keep_error_builds" does. In the END handler it > renames the build and install directories with a timestamp. Cleanup is > left to the user. Great. The machine has ample disk, so I'll add that setting.