raft/shard log errors

[Dieterbe]

while testing out #293 I noticed this in the daemon output:
https://gist.github.com/Dieterbe/9413295
right now:

~ ❯❯❯ ls -alh /tmp/influxdb/development/db/shard_db/00001/                                                                                    ⏎
total 28M
drwxr-xr-x 2 dieter dieter  180 Mar  7 11:14 .
drwxr--r-- 5 dieter dieter  100 Mar  7 10:22 ..
-rw-r--r-- 1 dieter dieter  26M Mar  7 10:23 000005.sst
-rw-r--r-- 1 dieter dieter 2.0M Mar  7 11:28 000006.log
-rw-r--r-- 1 dieter dieter 745K Mar  7 11:14 000007.sst
-rw-r--r-- 1 dieter dieter   16 Mar  7 10:22 CURRENT
-rw-r--r-- 1 dieter dieter    0 Mar  7 10:22 LOCK
-rw-r--r-- 1 dieter dieter  453 Mar  7 11:14 LOG
-rw-r--r-- 1 dieter dieter  64K Mar  7 10:22 MANIFEST-000002
~ ❯❯❯ 

although i did restart influx a few times in the meantime.

[jvshahid]

I think your tmp directory was nuked somehow. Did you do anything that could delete your tmp directory or is this something that you can easily reproduce ?

[Dieterbe]

i don't think i did anything to my /tmp
haven't seen it happen again, but if it happens again, anything specific i should look at?

[jvshahid]

no just ls tmp and make sure that the entire directory isn't gone. You can also check your cron to make sure you don't have a script that constantly clean up tmp.

I'll go ahead and close this issue, please reopen if you successfully repro this issue.

[Dieterbe]

note that most (if not all) common linux distributions implement the FHS (file hierarchy standard) which specifies that /tmp can be cleaned at any time.

quote man hier:

       /tmp   This directory contains temporary files which may be deleted with no notice, such as by a regular job or at system boot up.

[elcct]

A lot of distributions use tmpfs filesystem for /tmp point, which means all data will be gone after reboot