| Bug #62083 | ndbd unlink binary when specify BackupDataDir in cofiguration | ||
|---|---|---|---|
| Submitted: | 4 Aug 2011 13:45 | Modified: | 11 Nov 2016 23:36 |
| Reporter: | Andrey Lyubimov | Email Updates: | |
| Status: | No Feedback | Impact on me: | |
| Category: | MySQL Cluster: Cluster (NDB) storage engine | Severity: | S2 (Serious) |
| Version: | 7.1.15 | OS: | Linux (kernel 2.6.38-8 x86_64 (ubuntu 11.04)) |
| Assigned to: | MySQL Verification Team | CPU Architecture: | Any |
| Tags: | unlink binary | ||
[11 Oct 2016 23:36]
MySQL Verification Team
Hi Andrey can't reproduce on any of the modern systems, but can't reproduce on 7.1.15 :( If you still have some details about this bug (after this many years) you can send me the config file you used to get to this situation kind regards Bogdan Kecman
[12 Nov 2016 1:00]
Bugs System
No feedback was provided for this bug for over a month, so it is being suspended automatically. If you are able to provide the information that was originally requested, please do so and change the status of the bug back to "Open".
Description: The ndbd unlink it binary when kill the process. Situation repeats only when specific BackupDataDir presents in cluster configuration. When BackupDataDir not configured, process stops normally. # strace -f /usr/sbin/ndbd --- [cut] --- recvfrom(5, "result: ok\n\n", 512, MSG_PEEK, NULL, NULL) = 12 recvfrom(5, "result: ok\n", 11, 0, NULL, NULL) = 11 clock_gettime(CLOCK_MONOTONIC, {18795, 652367793}) = 0 clock_gettime(CLOCK_MONOTONIC, {18795, 652425385}) = 0 poll([{fd=5, events=POLLIN}], 1, 60000) = 1 ([{fd=5, revents=POLLIN}]) clock_gettime(CLOCK_MONOTONIC, {18795, 652545441}) = 0 recvfrom(5, "\n", 512, MSG_PEEK, NULL, NULL) = 1 recvfrom(5, "\n", 1, 0, NULL, NULL) = 1 close(5) = 0 close(6) = 0 close(5) = -1 EBADF (Bad file descriptor) unlink("/usr/sbin/ndbd") = 0 exit_group(0) = ? --- [/cut] --- When no BackupDataDir specifyed in config: # strace -f /usr/sbin/ndbd -- [cut] -- clock_gettime(CLOCK_MONOTONIC, {19440, 532508243}) = 0 recvfrom(5, "\n", 512, MSG_PEEK, NULL, NULL) = 1 recvfrom(5, "\n", 1, 0, NULL, NULL) = 1 close(5) = 0 close(6) = 0 close(5) = -1 EBADF (Bad file descriptor) unlink("") = -1 ENOENT (No such file or directory) exit_group(0) = ? Process 6743 detached -- [/cut] -- How to repeat: Start ndbd daemon when BackupDataDir specified in configuration for this ndb node. Send SIG_TERM to all ndbd process. Suggested fix: Possible PID file location failed in ndbd process and binary path used as PID file.