syslog is filled with error messages from OPAL_DUMP
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
ppc64-diag (Ubuntu) |
Fix Released
|
Undecided
|
Unassigned | ||
Trusty |
Fix Released
|
Undecided
|
Unassigned | ||
Utopic |
Won't Fix
|
Undecided
|
Unassigned |
Bug Description
-- Problem Description --
Steps to recreate the problem
1) Install Ubuntu 14.10 on P8 system - non virtualized mode
2) Install ppc64-diag and related packages.
When I examined /var/log/syslog, I observed it's filled with repeated messages show below.
Jul 22 09:00:37 lep8d OPAL_DUMP[108208]: Error accessing output dir: /var/log/dump (2: No such file or directory)
Jul 22 09:00:38 lep8d OPAL_DUMP[108210]: Error accessing output dir: /var/log/dump (2: No such file or directory)
Jul 22 09:00:39 lep8d OPAL_DUMP[108212]: Error accessing output dir: /var/log/dump (2: No such file or directory)
Jul 22 09:00:40 lep8d OPAL_DUMP[108214]: Error accessing output dir: /var/log/dump (2: No such file or directory)
I have not tried creating /var/log/dump directory to see if these stop. Can these repeated messages be suppressed ? It doesn't make sense to fill the syslog with continuous messages.
From the code in opal_errd/
#define DEFAULT_OUTPUT_DIR "/var/log/dump"
char *opt_output_dir = DEFAULT_OUTPUT_DIR;
...
rc = stat(opt_
if (rc != 0) {
}
We get the error because the file doesn't exist.
And since in
opal_errd/
we don't create file if it doesn't exist, the error will occure.
dir_fd = open(dirname(
Should I suggest to add the O_CREATE flag if it doesn't exist ?
Creating the directory manually gives the following output:
Jul 24 21:20:38 lep8d OPAL_DUMP[64949]: New platform dump available. File: /var/log/
Jul 24 21:20:38 lep8d kernel: [26142.017382] create_dump_obj: New platform dump. ID = 0x3a000000 Size 7256717
Jul 24 21:20:38 lep8d OPAL_DUMP[64947]: Failed to read platform dump: /sys/firmware/
Jul 24 21:20:38 lep8d OPAL_DUMP[64947]: Failed to acknowledge platform dump: /sys/firmware/
Jul 24 21:20:38 lep8d kernel: [26142.038146] dump_read_data: Extract dump failed for ID 0x3
Jul 24 21:20:39 lep8d OPAL_DUMP[64954]: Failed to read platform dump: /sys/firmware/
Jul 24 21:20:39 lep8d OPAL_DUMP[64955]: Failed to read platform dump: /sys/firmware/
Jul 24 21:20:39 lep8d OPAL_DUMP[64955]: Failed to acknowledge platform dump: /sys/firmware/
Jul 24 21:20:39 lep8d kernel: [26143.041011] dump_attr_read: Platform dump partially read.ID = 0x3a000000
Jul 24 21:20:39 lep8d kernel: [26143.041230] dump_read_data: Extract dump failed for ID 0x3a000000
root@lep8d:
dump image msglog update_flash
elog manage_flash sysparams validate_flash
root@lep8d:
initiate_dump
I think the ubuntu packaging should create this directory rather than relying on server to create it. This would be common with other daemons (which usually error out rather than create directories).
the other log messages around there are a bit of a worry too...
Related branches
tags: | added: architecture-ppc64le bugnameltc-113639 severity-high targetmilestone-inin1410 |
affects: | ubuntu → ppc64-diag (Ubuntu) |
Changed in ppc64-diag (Ubuntu): | |
status: | New → Confirmed |
tags: |
added: targetmilestone-inin1504 removed: targetmilestone-inin1410 |
Changed in ppc64-diag (Ubuntu Utopic): | |
status: | New → Won't Fix |
Changed in ppc64-diag (Ubuntu Trusty): | |
milestone: | none → ubuntu-14.04.3 |
------- Comment From <email address hidden> 2014-10-16 10:42 EDT-------
Looks like issue is not yet fixed in latest ppc64-diag version in Ubuntu 14.10 (v2.6.7)
-Vasant