Tumbleweed - iSCSI Initiator Yast Discovery/Connection Locks Up - stack smashing detected

English Network/Internet
1

I just installed the latest Tumbleweed (20191027) and yast is locking up while configuring a new connection (Network Services -> iSCSI Initiator -> Discovered Targets -> Connect). I am able to discover the targets and then after I specify the connection information (startup - automatic, authentication by targets) and and hit next (F10) Yast locks up. In the log files I can see a bunch of error messages from the iscisd service (detailed a little bit later).

The iSCSI target is on a QNAP NAS on my local LAN (and I have an OpenSuse Leap instance connected to another target on the same NAS).

I have tried this both on a clean install and in a virtual machine.

Via journalctrl I see the following messages:
Oct 30 10:38:54 tumbleweed-vm systemd[1]: Listening on Open-iSCSI iscsid Socket.
Oct 30 10:41:53 tumbleweed-vm iscsid[1146]: *** stack smashing detected ***: <unknown> terminated
Oct 30 10:41:53 tumbleweed-vm systemd[1]: iscsid.service: Main process exited, code=dumped, status=6/ABRT
Oct 30 10:41:53 tumbleweed-vm systemd[1]: iscsid.service: Failed with result ‘core-dump’.
Oct 30 10:41:53 tumbleweed-vm systemd[1]: iscsid.service: Service RestartSec=100ms expired, scheduling restart.
Oct 30 10:41:53 tumbleweed-vm systemd[1]: iscsid.service: Scheduled restart job, restart counter is at 1.
Oct 30 10:41:53 tumbleweed-vm iscsid[2613]: iscsid: Could not read data from db. Using default and currently negotiated values
Oct 30 10:41:53 tumbleweed-vm iscsid[2613]: iscsid: Invalid iscsi.FirstBurstLength of 0. Must be within 512 and 16777215. Setting to 262144
Oct 30 10:41:53 tumbleweed-vm iscsid[2613]: iscsid: Invalid iscsi.MaxBurstLength of 0. Must be within 512 and 16777215. Setting to 16776192
Oct 30 10:41:55 tumbleweed-vm iscsid[2613]: *** stack smashing detected ***: <unknown> terminated
(and this repeats over and over again)

Afterwards the service is in a failed state:
sudo systemctl status iscsid
iscsid.service - Open-iSCSI
Loaded: loaded (/usr/lib/systemd/system/iscsid.service; disabled; vendor preset: disabled)
Active: failed (Result: core-dump) since Wed 2019-10-30 11:07:06 EDT; 39s ago
Docs: man:iscsid(8)
man:iscsiuio(8)
man:iscsiadm(8)
Process: 6180 ExecStart=/sbin/iscsid -f (code=dumped, signal=ABRT)
Main PID: 6180 (code=dumped, signal=ABRT)
Status: “Syncing existing session(s)”

Oct 30 11:07:06 tumbleweed-vm systemd[1]: iscsid.service: Service RestartSec=100ms expired, scheduling restart.
Oct 30 11:07:06 tumbleweed-vm systemd[1]: iscsid.service: Scheduled restart job, restart counter is at 665.
Oct 30 11:07:06 tumbleweed-vm systemd[1]: Stopped Open-iSCSI.
Oct 30 11:07:06 tumbleweed-vm systemd[1]: iscsid.service: Start request repeated too quickly.
Oct 30 11:07:06 tumbleweed-vm systemd[1]: iscsid.service: Failed with result ‘core-dump’.
Oct 30 11:07:06 tumbleweed-vm systemd[1]: Failed to start Open-iSCSI.

I am then unable to start the iscsid service:> sudo systemctl start iscsid
Job for iscsid.service failed because a fatal signal was delivered causing the control process to dump core.
See “systemctl status iscsid.service” and “journalctl -xe” for details.

I had a previous version of Tumbleweed that was connected to this target and ran into this problem after reinstalling Tumbleweed the other day.

Any help would be appreciated.

Regards, Jeff

2

I’d suggest first verifying that your iscsi target is also running on system(s) that are fully updated and perhaps even running a version that is close to what you’re running in Tumbleweed.
I also don’t personally track what versions are running in Tumbleweed, since you’re already testing in a virtual machine, I’d recommend you also test in a LEAP machine.

You should submit a bug with your detailed findings to https://bugzilla.opensuse.org

The error suggests a coding error and not something a User would be expected to be able to address.

TSU

3

As an update to this issue - shortly, after posting this issue I submitted a bug (https://bugzilla.opensuse.org/show_bug.cgi?id=1155510). This ended up being an issue with iscisd: “new Authentication methods were added to the two that already existed, but one stack array, value_list in acl_init(), needed to be changed from [2] to [4], but was just changed to [3]”.

The iscsid changes were provided to me via an RPM which I tested and verified.

  • Jeff