and doesn't proceed to download any of the package lists via CLI (ssh). It updates fine from the LEDE web interface fine. What would cause that?
Also, if I try to d'load the latest trunk using wget (both https or http) it hangs for a couple minutes, and then downloads. Both ping and traceroutes look find to downloads.lede-project.org. I never had his issue with OpenWRT, so I am not sure why this is happening.
That delay of several seconds sounds strange. I noticed your comment yesterday on Openwrt forum and thought to write back, but didn't do that. I tested your exact command from my router and the download completion lasts about 1.8 seconds in total.
One guess for you:
I noticed that the download address first tries ipv6 and then ipv4 as a fallback. If you (or your ISP) has faulty half-working ipv6 config/routting/something, wget possibly first waits for ipv6 transfer and then falls back to ipv4. I remember seeing something with websites that resolve ipv6 address but the actual routing fails. You might test with ping
root@LEDE:~# time wget --no-check-certificate https://downloads.lede-project.org/releases/17.01.0/targets/ar71xx/generic/lede-17.01.0-r3205-59508e3-ar71xx-generic-tl-wdr4300-v1-squashfs-sysupgrade.bin
--2017-04-09 21:52:08-- https://downloads.lede-project.org/releases/17.01.0/targets/ar71xx/generic/lede-17.01.0-r3205-59508e3-ar71xx-generic-tl-wdr4300-v1-squashfs-sysupgrade.bin
Resolving downloads.lede-project.org... 2a01:4f8:202:43ea::3, 148.251.78.235
Connecting to downloads.lede-project.org|2a01:4f8:202:43ea::3|:443... connected.
...
lede-17.01.0-r3205-59508e3-ar71xx-generic-tl-wdr4300-v1-squashfs-sy 100%[=================================================================================================================================================================>] 3.38M 2.50MB/s in 1.3s
2017-04-09 21:52:09 (2.50 MB/s) - 'lede-17.01.0-r3205-59508e3-ar71xx-generic-tl-wdr4300-v1-squashfs-sysupgrade.bin' saved [3538948/3538948]
real 0m 1.77s
user 0m 0.24s
sys 0m 0.08s
IPv6 ping & traceroute:
root@LEDE:~# ping -6 downloads.lede-project.org
PING downloads.lede-project.org (2a01:4f8:202:43ea::3): 56 data bytes
64 bytes from 2a01:4f8:202:43ea::3: seq=0 ttl=55 time=32.510 ms
64 bytes from 2a01:4f8:202:43ea::3: seq=1 ttl=55 time=32.358 ms
^C
--- downloads.lede-project.org ping statistics ---
2 packets transmitted, 2 packets received, 0% packet loss
round-trip min/avg/max = 32.358/32.434/32.510 ms
root@LEDE:~# traceroute6 downloads.lede-project.org
traceroute to downloads.lede-project.org (2a01:4f8:202:43ea::3), 30 hops max, 64 byte packets
1 van1-er40.ip6.dnaip.fi (2001:14b8::6245) 1.924 ms 2.323 ms 14.578 ms
2 esp2-tr2.ip6.dnaip.fi (2001:14b8::9) 2.206 ms 2.449 ms 1.945 ms
...
But ... if I revert back to the latest OpenWRT snapshot, DESIGNATED DRIVER (Bleeding Edge, 50104), using sysupgrade the hang issue disappears without any network, dns, or firewall changes.
The only thing I have not tried yet is wiping LEDE clean with mtd -r erase rootfs_data, and restoring my basic network, samba, wireless, and other config from backups. Otherwise, it's an odd issue.
What about packages? Are there any incompatible packages from openwrt? I ask because I restore my packages via a backed up text file after setting up / activating extroot.
This method:
opkg update && for i in $(cat /etc/config/scripts/opkg_installed); do opkg install $i; done
There may naturally be some incompatible packages for any specific hardware & config combination, but it is difficult to see anything causing that kind of slowdown (especially with wget using http).
In general LEDE 17.01, LEDE master and Openwrt DD master are pretty much compatible regarding config files. I switch between the build every few days (due to my community builds for WNDR3700 and R7800),
One side note for you: hopefully you have pruned the opkg_installed list to contain only the packages that you really have selected and installed. If the list is unedited and contains also all packages that were dependencies for your selections, it is quite possible that you have installed some unnecessary packages (if they are not needed any more).
Reinstalling blindly all packages that were installed in the other system (Openwrt vs. LEDE) or (stable vs. snapshot) can be a bit dangerous. Not usually, but sometimes when there have been architecture changes.
Do you know of a good opkg_list with all the goodies in it? All major servers (vpn, samba asterisk Etc..), Luci, important network tools without all the bloat? I looked at mine, and I'm not sure what should stay, and what shouldn't.
[quote="kgoerbig, post:7, topic:2950"]
Do you know of a good opkg_list with all the goodies in it? All major servers (vpn, samba asterisk Etc..), Luci, important network tools without all the bloat? I looked at mine, and I'm not sure what should stay, and what shouldn't.
[/quote]All the goodies but without bloat. That is contradictory
It is highly personal what you need. So there is no generic opkg_list to be installed.
Your list includes e.g. all the default standard core packages like busybox, iptables, procd, ubus, ubox, firewall, swconfig etc. Those are always built-in. No need to consider re-installing them.
Additionally your list also contains packages that do not exist in the current LEDE ... like libpolarssl.
(And a bit strangely you install the ancient qos-scripts instead of sqm-scripts. But possibly that is intentional choice.)
That list looks rather bloated, to be honest
I build my own version of LEDE, so I edit directly the .config that defines the packages. My own list only includes the main packages that I want and those pull in the dependencies. E.g. Install luci-app-ddns and opkg automatically installs also "ddns" itself. Installing luci-ssl-openssl automatically installs also luci, uhttpd, libopenssl, ustreamssl-openssl etc.
I removed all the unnecessary packages from that list, and added the Luci packages from LEDE project page. Setup extroot, installed, rebooted. All is happy, on my my router now. No more hanging on d'loading sysupgrade.bin or opkg update.
