[ZBX-3159] vfs.fs.size - large filesystem and ZBX_NOTSUPPORTED, possible fix Created: 2010 Oct 28 Updated: 2017 May 30 Resolved: 2012 Jun 19 |
|
Status: | Closed |
Project: | ZABBIX BUGS AND ISSUES |
Component/s: | Agent (G) |
Affects Version/s: | 1.8.3 |
Fix Version/s: | 2.0.0 |
Type: | Incident report | Priority: | Major |
Reporter: | guilherme m. schroeder | Assignee: | Unassigned |
Resolution: | Fixed | Votes: | 6 |
Labels: | agent | ||
Remaining Estimate: | Not Specified | ||
Time Spent: | Not Specified | ||
Original Estimate: | Not Specified | ||
Environment: |
All Linux 32/64bits |
Issue Links: |
|
Description |
Trying to monitor a large filesystem returns ZBX_NOTSUPPORTED.
From df -h: Stracing it, we see that statfs() gets EOVERFLOW: statfs("/var/run/sr-mount/280465a6-b940-5cf9-f9b0-141b5090f274", 0xbfe206ec) = -1 EOVERFLOW (Value too large for defined data type) The solution i found was to recompile the client using this:
Then it works fine, because it uses statvfs64():
From strace: stat64("/var/run/sr-mount/280465a6-b940-5cf9-f9b0-141b5090f274", {st_mode=S_IFDIR|0755, st_size=249, ...}) = 0 I've been playing with zabbix-1.8.3/src/libs/zbxsysinfo/linux/diskspace.c, changing from statvfs/statfs to statvfs64/statfs64, but got into problems compiling and maybe it's not the better way to do this. On 32 or 64 bits archs, zabbix_agentd uses statfs/statvfs and should use statfs64/statvfs64, so people with big file systems don't get ZBX_NOTSUPPORTED. |
Comments |
Comment by Aleksandrs Saveljevs [ 2010 Oct 29 ] |
While we are at it, maybe we should also try to fix |
Comment by richlv [ 2010 Dec 08 ] |
probably the same as |
Comment by Gertjan Awater [ 2011 Nov 17 ] |
meanwhile, what is the largest filesystem possible without getting ZBX_NOTSUPPORTED ? |
Comment by Alexander Vladishev [ 2012 Jun 19 ] |
Fixed with |