Pydio not starting / installing

I have a recurring problem that the cells installation either fails during the final step or if it succeds, cells won’t start properly. Some services will start, but many won’t, failing with the message runtime/cgo: pthread_create failed: Resource temporarily unavailable. Sometimes this will result in the start command to fully crash, sometimes only some services crash. If I get a running instance and login, I get an Internal Server Error-Popup at the bottom and can’t do anything useful beside looking at the server settings.

All I found regarding this error that it has to do with process / thread limitations, which I don’t see on my system. Here are the (I think) important stats:

I’m trying to run cells as a dedicated user “pydio” on a Linux V-Server running Ubuntu 18.04 LTS and Plesk Onyx as a managing frontend.

$ uname -a
Linux <redacted> 4.15.0 #1 SMP Thu Apr 11 18:11:44 MSK 2019 x86_64 x86_64 x86_64 GNU/Linux
$ ./cells version
Warning: no private IP detected for binding broker. Will bind to <redacted>, which may give public access to the broker.

Pydio Cells Home Edition (1.5.2)
Published on 07 May 19 15:06 +0000
Revision number 4ae5255e1b28b01d9dea14ec1b386069194aa547
$ cat /proc/sys/kernel/pid_max
32768
$ cat /proc/sys/kernel/threads-max
3092863
$ ulimit -Sa
core file size          (blocks, -c) 0
data seg size           (kbytes, -d) unlimited
scheduling priority             (-e) 0
file size               (blocks, -f) unlimited
pending signals                 (-i) 1546431
max locked memory       (kbytes, -l) 16384
max memory size         (kbytes, -m) unlimited
open files                      (-n) 65536
pipe size            (512 bytes, -p) 8
POSIX message queues     (bytes, -q) 819200
real-time priority              (-r) 0
stack size              (kbytes, -s) 8192
cpu time               (seconds, -t) unlimited
max user processes              (-u) 62987
virtual memory          (kbytes, -v) unlimited
file locks                      (-x) unlimited
$ ulimit -Ha
core file size          (blocks, -c) unlimited
data seg size           (kbytes, -d) unlimited
scheduling priority             (-e) 0
file size               (blocks, -f) unlimited
pending signals                 (-i) 1546431
max locked memory       (kbytes, -l) 16384
max memory size         (kbytes, -m) unlimited
open files                      (-n) 1048576
pipe size            (512 bytes, -p) 8
POSIX message queues     (bytes, -q) 819200
real-time priority              (-r) 0
stack size              (kbytes, -s) unlimited
cpu time               (seconds, -t) unlimited
max user processes              (-u) 62987
virtual memory          (kbytes, -v) unlimited
file locks                      (-x) unlimited
$ mysqld --version
mysqld  Ver 10.3.11-MariaDB-1:10.3.11+maria~bionic for debian-linux-gnu on x86_64 (mariadb.org binary distribution)

Memory usage is at about 800MiB/8GiB, no swap available and only ~60 processes running when starting cells. Here are two example outputs [1] [2] that I encounter.

Any help or clue on how to gather more information or solve this issue is greatly welcome.

Hi cshark,

Can you please try to build the binary directly on your server ?

Thanks,
Greg

I built it directly from the repository following the guide, but no change. Fresh installation fails with the same error, again shortly after starting pydio.grpc.data.objects with the local1-source. Running the go tests also showed nothing particular, except for two errors:

=== RUN   TestBinaryStoreHandler_ReadNode

  Test Read Store Node, no Data Source info available 2019-05-18T16:49:29.403+0200      ERROR    Could not find DataSource store {"currentSources": []}
✔✔✔


166 total assertions

--- PASS: TestBinaryStoreHandler_ReadNode (42.00s)

<snip>

=== RUN   TestBoltStore_CRUD

  Test CRUD ✔✔✔✔✔✔✔✔✔✔✔✔✔2019-05-18T16:49:05.015+0200        ERROR    listVersions
    {"error": "{\"id\":\"versions\",\"code\":404,\"detail\":\"No bucket found for this node\",\"status\":\"Not Found\"}"}
✔


20 total assertions


  Test DeleteVersionsForNode ✔✔✔✔✔


25 total assertions

--- PASS: TestBoltStore_CRUD (0.13s)

From your logs it looks like a cgo runtime error. We don’t actually use cgo in our codebase but we rely on libraries that do use it. I’ll try to build a version with CGO_ENABLED=0 to see if that changes anything. Your server config looks ok and I can’t see anywhere where we would reach a server limit with those settings so let’s start off with that build.

Thanks,
Greg

Any news on this topic?