[mythtv-users] rpm build of master: sudden new failure

[John Pilkington]

On 04/08/2019 23:20, Bill Meek wrote:
> On 8/4/19 4:04 PM, John Pilkington wrote:
>> Last week (and several times in the past month) I built rpms for 
>> master for both Fedora 29 and el7.  The builds were done in F29 using 
>> Gary B's script with my normal dropping of libvpx for el7, and both 
>> mythtv systems are running well.
>>
>> Now, two commits later, the build fails.  The apparent reason is 'No 
>> space on device.'  This is presumably /tmp, which shows as 2 GB; 
>> system RAM is 4 GB.  I tried a google-suggested mount -o increase to 4 
>> GB, and also tried reverting to kernel 5.1.18 but still see the same 
>> problem. ubuntu PPA shows a build at the same commit.
>>
>> Attached are the end of root.log and build.log
>>
>> root.log shows an error (can't access /var/lib/mock) but with a 
>> cryptic comment and is probably ok.   I don't think I have a saved 
>> good version.
>>
>> build.log has several 'Failed to write block' and then 'No space on 
>> device,' again with a /var reference.   I tried deleting a few old 
>> recordings but saw no change.
>>
>> Ideas welcome.
>>
>> John P
> 
> The MythTV buildbot just completed builds on: master-f29-64bit and 
> master-f28-64bit
> a few minutes ago (and roughly 100 prior to that.) It doesn't appear to 
> be a commit
> issue.
> 
> *buntu user here, so I don't know if Fedora's /var/tmp is a link to 
> /tmp, but /var/tmp
> appears to be where the space issue is. df --si /var/tmp. I read the 
> comment to be
> about btrfs only, not including /var/tmp

Thanks Bill, that looks helpful.  I couldn't see where builddir really 
was, but this history is suggestive.  I'll see what I can purge.

John

{{{

konsole tab output while BUILD_MYTHTV_RPMS was attempting to build 
31Pre546, perhaps 50 minutes in, then at 'random' times until it failed 
again.

[john at HP_Fed rpm]$ df --si /var/tmp
Filesystem               Size  Used Avail Use% Mounted on
/dev/mapper/fedora-root   53G   50G  804M  99% /
[john at HP_Fed rpm]$ df --si /tmp
Filesystem      Size  Used Avail Use% Mounted on
tmpfs           2.1G  309M  1.8G  16% /tmp
[john at HP_Fed rpm]$ df --si /var/tmp
Filesystem               Size  Used Avail Use% Mounted on
/dev/mapper/fedora-root   53G   50G  216M 100% /
[john at HP_Fed rpm]$ df --si /var/tmp
Filesystem               Size  Used Avail Use% Mounted on
/dev/mapper/fedora-root   53G   50G  129M 100% /
[john at HP_Fed rpm]$ df --si /var/tmp
Filesystem               Size  Used Avail Use% Mounted on
/dev/mapper/fedora-root   53G   50G  691M  99% /

That may have been during the clean-up phase after failure

and this is back to the quiescent state...

[john at HP_Fed rpm]$ df --si /var/tmp
Filesystem               Size  Used Avail Use% Mounted on
/dev/mapper/fedora-root   53G   45G  5.7G  89% /
[john at HP_Fed rpm]$ df --si /var/tmp
Filesystem               Size  Used Avail Use% Mounted on
/dev/mapper/fedora-root   53G   45G  5.7G  89% /
[john at HP_Fed rpm]$

}}}