-
-
Notifications
You must be signed in to change notification settings - Fork 229
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
unable to start default pelias docker due to "Elasticsearch service did not come up, check configuration" error #214
Comments
Can you please post the output of these commands: sudo netstat -plnt docker ps -a | grep pelias |
Output of netstat:
Output of docker ps:
|
Hi @wkulesza, The output of The output can be quite verbose so can you please paste it all here and be sure it's formatted correctly? Your copy/paste above appears to have lost quite a few newlines making it hard to read. For example on my machine,
|
I fixed the formatting above, the trick for markdown is to paste it between a pair of triple backticks |
Thanks @missinglink for helping with formatting. Pasting in the elasticsearch log - obviously is very long, but i think this part is repeating towards the end:
|
This directory should be available on your host machine at Is there anything unusual about that directory? ie. permissions or network mount etc? |
Possibly related:
I suspect the issue is related to user permissions. |
@missinglink DATA DIR is set in portland-metro to ./data so i understand this directory is /home/wojciech/pelias/docker/projects/portland-metro/data/elasticsearch ? and not usr/share/elasticsearch ? that elasticsearch in home/wojciech is empty |
@wkulesza the ownership of the directory is likely the issue, it can't be owned by root. This is a common problem, and we were working on some better automatic detection/fixing of this issue in #74, maybe we will pick that up. For now I happen to have just listed some basic troubleshooting tips in another comment, let me know if any of these work for you: #31 (comment) |
@orangejulius will work on it, but for the time being, when i try to run it as normal user (without sudo) i get permission errors for creating symbolic link (which is normal, as standard user can't work on /usr/local/bin directory)
|
While other `pelias` commands should not be run with `sudo`, `sudo` will generally be required to create a symbolic link in `/usr/local/bin/pelias` as described in our quickstart script. This change adds `sudo` to that command and clarifies that it's the only part of the quickstart script that should use `sudo`. Connects #214
Right. Only the |
OK. i made that manually now, added ln command as sudo
|
Is |
@orangejulius yes, /usr/local/bin is part of my PATH, i checked this. I also tried to manually set the ln first (as sudo user) and then perform the pelias.sh script with that ln line commented out. even after going sudo, i get no response from pelias command:
|
Just to add some notes, i was now trying to run pelias compose pull as normal user but got a lot of permission denied (all folders in /home/wojciech/pelias are owned by that user (wojciech)
|
I don’t know if I should provide more details but as root I went in to Wlkp projects folder and was able to perform all pelias commands including pelias elastic wait - which finished with success! But running instance doesn’t return addresses to inquiries. |
From my last observatio0n @missinglink @orangejulius docker with elasticsearch is constantly restarting - once every 1 minute it's going up, then after 1-3 secs it's restarting again. |
The directory is owned by root and 755 then no user other than root can write to it. |
@missinglink Hi, thanx for pointers. I chmoded portland-metro to 777 and this allowed me to go forward (elasticsearch is up) but stumbled upon this upon pelias elastic create for openaddresses: Error: dir option must be relative to "/tmp", found "data/openaddresses". I have not set tmp as a data folder, but projects/portland-metro/data/ Any pointers ? |
Agh nice catch, this is a bug introduced last week by our dependency management bot. I've reverted that commit in pelias/openaddresses#473 which has automatically published a new docker image. You can pull the new image with I'm going to have to go through the other repositories and check them for this same error, so let me know if you see it again. |
Same problem here, not solved with latest docker images. Fedora Server 32. |
Hi folks, we've just merged some improvements in #220 that should help with this sort of problem. The If you get a chance please try things out with the latest code. We recommend removing your data directory completely to ensure all permissions are set properly going forward. |
Just to follow up, I successfully run through this updated version without having to be root (just needed to provide sudo pasword for ln command), but ended up with pelias test run that provided me with this error: my docker ps says everythjing is running and I haven't noticed any errors, when following the standard tutorial: curl also says: connect econnrefused my docker ps says:
edit1: sorry, tagging @missinglink and @orangejulius - i didn't want to open a new issue, as this is a continuation. and also, i tried with previous project (only one region in poland), just in case, 8gb of ram wouldn't be enough for "poland" project, but ended up with the very same error connect ECONNREFUSED 172.22.0.4:4400 edit2: i noticed that libpostal and interpolation do restart according to status of docker ps:
edit3:
|
Hi @wkulesza libpostal alone needs ~2GB of RAM to load its models. It's used by the The "Cannot allocate memory" error seems to indicate that your machine doesn't have adequate memory to run Pelias, our documentation seems to indicate that 8GB is sufficient but that doesn't seem to be correct. Adding RAM would be a simple solution, otherwise you can run Pelias without the interpolation service, although 8GB will likely not yield great performance from There's actually a long standing PR to make |
Describe the bug
Unable to start default pelias docker with default portland-metro project due to "Elasticsearch service did not come up, check configuration" error
Steps to Reproduce
Creating pelias _elasticsearch .. done
waiting for elasticsearch service to come up
...........................
Elasticsearch service did not come up, check configuration
Creating pelias_schema_run ... done
Create index
Elasticsearch ERROR: 2020-09-15T11:53:21Z
Error: Request error, retrying
GET http://elasticsearch:9200/ => connect ECONNREFUSED 172.19.0.2:9200
[..]
Elasticsearch WARNING: 2020-09-15T11:53:21Z
Unable to revive connection http://elasticsearch:9200/
Elasticsearch WARNING: 2020-09-15T11:53:21Z
No living connections
[...]
unsupported elasticsearch version. try: >=7.4.2
...
all next log items seem to be ok apart from tests.
Expected behavior
Elasticsearch should load and website api should return correct results?
Environment (please complete the following information):
Ubuntu Ubuntu 20.04 LTS Server 64bit
Docker version 19.03.12, build 48a66213fe
docker-compose version 1.27.2, build 18f557f9
Pastebin/Screenshots
Api works and returns data, but incomplete
[http://prntscr.com/uhnaf4]
References
As requested by @missinglink in issue #180 180 adding this a separate issue.
The text was updated successfully, but these errors were encountered: