serverfault.com

I use lvm on my raspberry pi 5 running Alpine 3.23 with internal nvme, I also have two external USB disks (WD 1Tb and 6TB). All three disks are in the same volume group vg0 with nvme: lv_root for / lvol0 for /home, and lvol1 for /var/lib 6TB HDD: one lv named lvol2, size 930GB 1TB HDD: empty Problem lvol2 on the external hdd is not mounted during boot despite being listed correctly in fstab. The other lvs mount (all three are on the nvme disk). Workaround: manually mount lvol2 after boot and restart affected services. Context fstab (relevant lines only); lvol2 has UUID c06ad240-9171-403b-8b76-71ec7d25cb46: /dev/vg0/lv_root / ext4 rw,relatime 0 1 UUID=E8F3-9212 /boot vfat rw,re

I'm debugging a very strange issue on a Proxmox server (running on Debian 12). It is connected via a 10G link (SFP+ DAC) to a switch. Another 10G link goes to another very similar server: SrvA -- 10G DAC --> Switch -- 10G DAC --> SrvB 10.5.1.2 10.5.0.102 10.5.8.2 When I ping SrvA from SrvB, all looks normal, with RTTs well below 1ms and a normal 1s interval. But if I flip the ping direction and ping SrvB from SrvA (as indicated by the arrows above), strange things happen. Most pings work just fine, also with RTT<1ms. But every about 20s, ping will just stall. However, it is not the normal the-response-is-delayed thing, but instead apparently a delay on the sending side, either before the request is even sent, or after the response is received and before the data is printed. After a few seconds, the line is printed with a normal RTT, with no gap in the sequences, no offending TTL, all normal. The only indication

We are fond of desktop shortcuts. We create colorful ico to distinguish each of the shortcuts. But lately, all new desktop shortcuts we created are showing as only white panel, the ico graphics are missing. Yes we already tried, enabled GPO "Allow the use of remote paths in file shortcut icons" in "Computer Configuration → Administrative Templates → Windows Components → File Explorer" rebuilt icon cache

I am currently trying to configure Cinder + iSCSI + NetApp in VHI (OpenStack) using the NetApp Generic Driver, with the goal of providing external storage to instances for their volumes. I have reviewed the following documentation: NetApp – OpenStack Deploy & Operations Guide OpenStack – Cinder NetApp Volume Driver Documentation Virtuozzo Hybrid Infrastructure – Admin Guide (Attaching External iSCSI Storage) When configuring Cinder, the following parameters must be defined: -volume_driver = cinder.volume.drivers.netapp.common.NetAppDriver -netapp_storage_family = ontap_cluster -netapp_storage_protocol = iscsi -netapp_vserver = openstack-vserver -netapp_server_hostname = myhostname -netapp_server_port = port -netapp_login = username -netapp_password = password However, I have several questions regarding the required configuration on the NetApp side. From what I understand, the Driver actually connects

When I install Docker on a new machine, I get a version that is newer than 28.1.1. E.g. recently, the ubuntu-latest vmImage in Azure Pipelines updated Docker to 29.1.5. Then, when I login to our private registry (running the registry:2 image): echo '$(DOCKER_PASSWORD)' | docker login -u=$(DOCKER_USER) --password-stdin $(DOCKER_REGISTRY), I get (as expected): "Login Succeeded". But now, when I try to pull an image: docker image pull $(DOCKER_REGISTRY)/$(DOCKER_USER)/package, I get: Error response from daemon: unknown: failed to resolve reference "//package:latest": unexpected status from HEAD request to https:///v2//package/manifests/latest: 401 Unauthorized Apparently, the Docker client "forgets" to include the Authorization header. By experiment, I found that this changed afte

In the from section, on the bottom of the email confirmation, it says "Your Project-342579022 Team". Is there an easy way to change the name from Project-9859283 to name like "Support" or "Customer Support", from within my console? This way, it would read "Your Support Team". Thanks!

I’m designing a system that is logically symmetric at rest, but where failure and recovery break that symmetry. At steady state, components are interchangeable and no node is “special.” However, during failures (shutdowns, partial outages, recovery), behavior becomes asymmetric: ordering, locality, or timing suddenly matters. This raises a design question: • If a system is symmetric in its idealized model, but failure introduces asymmetry, how should that be handled at the architecture level? • Is it better to: • Explicitly introduce a coordinator / anchor to manage failure and recovery? • Preserve symmetry and accept that recovery paths will implicitly break it? • Or redesign so that failure modes are first-class and symmetry is only approximate? I’m less interested in implementation details and more in principled guidance from reliability engineering or distributed systems practice: how experienced operators think about symmetry vs.

I have an old HP Proliant DL385 G1 server. When powered on it hangs at the HP Proliant startup logo screen. All 6 drives power up and show a green disk drive light. not sure what to check next.

Our e-mail server (FreeBSD-14.4) runs sendmail and is using SpamAssassin: INPUT_MAIL_FILTER(`spamassassin',`S=local:/var/run/spamass-milter.sock, F=T, T=C:15m;S:4m;R:4m;E:10m')dnl As I noticed, even the locally-originating emails -- such as those from the local cron-jobs -- are routed through the milter, which is silly (and wasteful as some tests are heavy)... In an effort to exempt the local machine -- and the LAN -- I added the following entries to the access-map: Spam:192.168 SPAM_FRIEND Spam:127.0.0.1 SPAM_FRIEND Spam:localhost SPAM_FRIEND but that had no effect. What is the right course of action?

I was wondering if anyone could help me understand how to efficiently take control of an NTFS share that is restored from a backup and manage permissions on it. The share is rather large and has 10+ levels of directories with varying permissions and inheritance settings. I don't want to wipe all of the permissions but I'm working on the backup on another machine and the account has no permissions to the folder set. I would like to take ownership of the entire directory tree and allow the account I'm working under to have full control to the folder set for managing of folder permissions. Is there an easy way to do this? Right now I'm stuck at taking ownership of one level (recurse fails because there's no permissions to view the folder contents). After that, adding the account I'm using to the permissions with full control and then repeating the process on the next level with every folder. This seems impractical and I'm not sure how to do this efficiently.