Resource temporarily unavailable

[shibumi]

Expected behavior

I should be able to rebuild and update the cache (openvasmd --update --progress && openvasmd --rebuild --progress) and start the openvas-scanner.

Current behavior

I am not able to update or rebuild the cache via openvasmd nor I am able to restart the openvas-scanner.

When I try to update or rebuild the cache the whole process just hangs. When I attach strace to the process I get the following messages in an infinite loop:

select(17, [16], [], NULL, {tv_sec=0, tv_usec=822418}) = 0 (Timeout)
recvfrom(16, 0x7ffcaa7d20ff, 1, MSG_PEEK, NULL, NULL) = -1 EAGAIN (Resource temporarily unavailable)
select(17, [16], [], NULL, {tv_sec=1, tv_usec=0}) = 0 (Timeout)
recvfrom(16, 0x7ffcaa7d20ff, 1, MSG_PEEK, NULL, NULL) = -1 EAGAIN (Resource temporarily unavailable)
select(17, [16], [], NULL, {tv_sec=1, tv_usec=0}) = 0 (Timeout)
recvfrom(16, 0x7ffcaa7d20ff, 1, MSG_PEEK, NULL, NULL) = -1 EAGAIN (Resource temporarily unavailable)
select(17, [16], [], NULL, {tv_sec=1, tv_usec=0}) = 0 (Timeout)
recvfrom(16, 0x7ffcaa7d20ff, 1, MSG_PEEK, NULL, NULL) = -1 EAGAIN (Resource temporarily unavailable)

I have no idea what is happening here. In the logs I can find the following events in the openvasmd.log:

The first bad message is this one here:

md manage:WARNING:2019-02-26 18h07.15 utc:13695: sql_exec_internal: sqlite3_step failed: cannot start a transaction within a transaction
md manage:WARNING:2019-02-26 18h07.15 utc:13695: sqlv: sql_exec_internal failed
[...]
md manage:WARNING:2019-02-26 18h07.15 utc:13695: sql_exec_internal: sqlite3_step failed: cannot start a transaction within a transaction
md manage:WARNING:2019-02-26 18h07.15 utc:13695: sqlv: sql_exec_internal failed
md manage:WARNING:2019-02-26 18h07.15 utc:13695: manage_schedule: manage_update_nvti_cache error (Perhaps the db went down?)
md manage:WARNING:2019-02-26 18h07.25 utc:13695: sql_exec_internal: sqlite3_step failed: cannot start a transaction within a transaction
md manage:WARNING:2019-02-26 18h07.25 utc:13695: sqlv: sql_exec_internal failed
[...]

Today I have restarted the service and it looks like this:

d   main:MESSAGE:2019-02-27 11h06.19 utc:2660:    OpenVAS Manager version 7.0.3 (DB revision 184)
md manage:   INFO:2019-02-27 11h06.19 utc:2660:    Getting users.
md   main:MESSAGE:2019-02-27 11h06.25 utc:2661:    OpenVAS Manager version 7.0.3 (DB revision 184)
md   main:   INFO:2019-02-27 11h06.25 utc:2661: rebuild_nvt_cache_retry: Reloading NVT cache
md   main:   INFO:2019-02-27 11h06.25 utc:2662: update_or_rebuild_nvt_cache: Updating NVT cache
base gpgme:MESSAGE:2019-02-27 11h06.25 utc:2662: Setting GnuPG dir to '/var/lib/openvas/openvasmd/gnupg'
base gpgme:MESSAGE:2019-02-27 11h06.25 utc:2662: Using OpenPGP engine version '2.2.4'
md   main:   INFO:2019-02-27 11h06.25 utc:2662:    Updating NVT cache.

When I try to start the scanner it just times out without any error message in the logs.

Steps to reproduce

1. Install OpenVAS on Ubuntu 18.04 via PPA
2. do a few scans
3. wait a few days.. (sorry no Idea how to reproduce this :( )

GVM versions

gsa: 7.0.3

gvm: 7.0.3

openvas-scanner: 5.1.3

gvm-libs: 9.0.3

openvas-smb: not existent?

Environment

Operating system: Ubuntu 18.04

Installation method / source: (packages, source installation) PPA

Logfiles

See logs above

[bjoernricks]

Looks like it is related to #347

[cfi-gb]

If the rebuild of the openvassd "hangs" then it is more likely an insufficient redis-configuration like outlined in https://community.greenbone.net/t/hint-redis-setup-configuration-for-gse-gvm-openvas/333

[bjoernricks]

Should be re-tested with current version. If it is still valid please re-open.