Lancache Epic and RIOT (probably some others too, steam is fine) downloading keep doing start and stop, start and pause etc. #192
Comments
coddec
changed the title
|
For record. |
|
I second this issue, same situation right now. |
|
Tried every fix I can think of. None worked. |
|
also having this issue |
Disabled nginx slicing on my end i had little improvements but its still not reliable as it used to be unfortunately. |
|
Do they have any version tags to use other than latest? Edit1: here the tag i used: Edit2: |
|
Where are you all located who are having this issue. We've been running the latest version of lancache all weekend at stratlan (~150 clients) and haven't seen this issue once. I'm wondering if there's a regional issue on one of epic's upstream. The errors you are seeing implies that the upstream server is disconnecting from lancache before it's finished downloading and the epic client is doing a 10s walkback before trying again. |
|
Frankfurt, Germany. |
|
Australia |
|
I'm wondering if this is CDN's (badly) handling capacity and attempting to throttle lancaches. What event size are you guys running as well Note for future me> Investigate proxy_next_upstream_tries |
|
Only one client in my LAN. |
|
Only two clients on LAN for now, too. Planning a party in two weeks I was also having this issue with Steam downloads, but I've since forced rerun the prefill script and it filled/redownloaded around 10% of the library (not entire games, only partial) which i found interesting. Since then I haven't encountered any issues with Steam. Haven't had the time to test other publishers. But will try and confirm the continuation of this issue in the comming days. |
|
I've noticed this on Riot launcher for almost a year now. I'm located in Finland. |
|
Same problem here with RIOT client since last two months |
|
Have been having the same issue with Riot games for at least 2 years. Atrocious download speed, 0kbps most of the time. I had to stop using lan cache altogether. I never saw an issue about it open so I figured something was wrong on my end. (I'm in Montreal, Canada if that makes a difference) |
|
Can i stop caching riot games? |
|
I recommend that you do this |
|
Linking this issue which has some additional discussion : uklans/cache-domains#209 |
|
I experienced the same issue last weekend "We're sorry. It looks like we're having trouble connecting. We'll continue to retry" on my little LAN Party with 4 Clients. |
Describe the issue you are having
Case 1
Without lancache, Epic downloads GTA fine with full WAN speed
Case 2
With lancache, Epic prefill (https://github.com/tpill90/steam-lancache-prefill) for GTA will download at full WAN speed no stop or pause.
Case 3
With lancache, Epic and RIOT games downloading via Epic client or RIOT client will download at full speed at start then drop to 0, stay at 0 for sometime then back to full speed, then repeat.
While this is happening lancache will have error log entries like shown below
(I can see that the container is not downloading too, certainly not that container is downloading the client is waiting case. In other words not this https://lancache.net/docs/common-issues/#slow-download-speeds)
The Epic client also show error message "We're sorry. It looks like we're having trouble connecting. We'll continue to retry" Click on it lead to https://www.epicgames.com/help/en-US/c-Category_EpicGamesStore/c-EpicGamesStore_ErrorCodes/is-0005-unable-to-connect-a000084850 stating "IS-0005: Unable to connect, This error is caused when your computer is not connected to the internet."
My internet is certainly working and there's no fluctuation, otherwise Epic prefill will experience the same issue.
Client computer to LAN and internet connection is also working fine.
Content are cached though. Enough disk space, enough CPU and RAM etc. Client PC is fine too.
12MB/s should not upset NIC since tested on 1g NIC and 10g sfp+ dac.
This issue makes the download taking way longer than it should be.
Case 4
Steam download works fine. Content are cached.
Tried to clear all cache
Tried to change slice size between 1M - 8M
Tried to disable slice, by setting to 0
Add multiple IP to lancache host tested, no help
No client RAM, CPU, Disk issue
No lancache host RAM, CPU, Disk issue
...
How are you running the container(s)
DNS Configuration
Output of container(s)
Any help appreciated.
TIA
The text was updated successfully, but these errors were encountered: