Unable to Update WP Plugins -cURL timeout error
I was using Digital Ocean for almost 3 years since their launch of Indian DC. When Linode launched Mumbai DC, I did some benchmarks and found that it has a great CPU compared to Digital Ocean and handles a lot of processes.
Within one month of Mumbai DC launch by Linode, I migrated my websites and even 4-5 clients to Linode. Initially, there weren't many issues but some temporary network glitches but in recent 3-4 months, the network service has become worse.
I manage around 30+ Indian sites from 5-6 Linode clients and they all face the same problem which is cURL timeout which makes it extremely hard to update WordPress plugins/themes. Most of the time plugins update fails or it takes huge time to update the plugin.
I still have 2-3 clients who use Digital Ocean and with the same server setup their WordPress sites never fail during the plugin/theme update and we never get cURL timeouts.
I tried raising the same issue from different Linode accounts and never got a satisfying reply. All they do ask for MTR reports and even after providing everything the Linode support asked, the outcome was Zero improvement.
We can manage with fewer requests handling but we don't want to experience sites stucking in maintenance mode due to the networking issues. I hope you will improve networking in the next few years and maybe, I will switch back to Linode then.
It hasn't been a nice experience at all :(
Does anyone around here face a similar issue, especially, with Mumbai data Center?
I have been using Linode for a year and recently faced the same issues with my server too. As we know recently WordPress 5.5 was released and I tried to update my websites (which are hosted on Linode Mumbai DC) but I was getting time out and maintenance mode error in all sites. Tried to contact support but they said everything is fine from our side. So I'm writing here as if we can get some attention from the Linode team.
All they do ask for MTR reports and even after providing everything the Linode support asked, the outcome was Zero improvement.
What actions did the support team take?
Are they aware there has been no improvement?
@andysh When I submitted the MTR reports from almost 2-3 accounts, they told me the following:
The MTR report suggests that there is a routing loop outside of our network. Unfortunately, there is hardly anything that we can do to fix the outside network.
There were aware that the problem isn't fixed but they always sound like it's not the Linode problem but the other routing networks.
I am not a networking engineer myself but I simply tried having the same setup with Digital Ocean, Bangalore and never faced the cURL Timeout and WordPress update issue.
Also, it's not like I follow the same server setup for every website, I have been using RunCloud and similar Cloud VPS Management Tools which powers the thousands of sites and servers. Still, the problem related to cURL Timeouts was there on Linode servers.
Thank you all for these additional data points and sharing your experience here. We hear your frustration with the degraded connectivity to Wordpress, and our engineers are aware of and actively working to resolve this issue. This is difficult because, as @nsgoyat correctly surmised, this problem does not arise on our network. We’ve reached out to some other ISPs in the hopes that they will be able to circumvent the bad paths here.
I wanted to chime in with a suggestion that might meet your needs while our network engineers work toward a potential resolution.
I spun up a test Linode in Mumbai, and it seems as if connectivity to Wordpress’ gitub repository might be significantly faster than using wordpress.org:
time git clone https://github.com/Wordpress/Wordpress.git Cloning into ‘Wordpress’... remote: Enumerating objects: 44, done. remote: Counting objects: 100% (44/44), done. remote: Compressing objects: 100% (35/35), done. remote: Total 329698 (delta 29), reused 16 (delta 9), pack-reused 329654 Receiving objects: 100% (329698/329698), 230.35 MiB | 1.08 MiB/s, done. mtr -rwzbc100 github.com Start: 2020-08-27T08:26:25-0400 HOST: localhost Loss% Snt Last Avg Best Wrst StDev 1. AS55429 gw-li1216.linode.com (18.104.22.168) 0.0% 100 0.2 0.2 0.1 0.3 0.0 2. AS??? 10.214.64.8 0.0% 100 0.3 0.4 0.2 2.3 0.4 3. AS??? 10.214.32.52 0.0% 100 13.1 2.4 0.8 21.5 3.4 4. AS??? 10.214.32.3 0.0% 100 1.4 1.0 0.3 16.1 2.2 5. AS63949 22.214.171.124 0.0% 100 0.4 0.9 0.3 10.3 1.1 6. AS33843 100.76.112.61 0.0% 100 0.8 1.1 0.4 38.2 3.8 7. AS??? 172.23.78.238 0.0% 100 1.1 1.9 0.8 18.5 2.7 8. AS4755 126.96.36.199.static.Mumbai.vsnl.net.in (188.8.131.52) 0.0% 100 0.9 1.2 0.8 6.4 0.8 9. AS16509 184.108.40.206 0.0% 100 1.7 2.5 1.5 12.8 1.8 10. AS16509 220.127.116.11 0.0% 100 1.6 2.0 0.9 17.6 2.4 11. AS16509 18.104.22.168 0.0% 100 5.8 4.2 3.0 17.9 2.3 12. AS16509 22.214.171.124 0.0% 100 4.5 7.3 2.4 40.3 7.7 13. AS16509 126.96.36.199 0.0% 100 2.4 7.3 2.3 40.7 9.6 14. AS??? ??? 100.0 100 0.0 0.0 0.0 0.0 0.0
I hope that this helps, and I’m sorry that your experience on our platform isn’t currently meeting your expectations. Threads like these are really valuable. It’s really helpful to know where we’ve missed the mark so we can work on making things better.
Got a response regarding my MTR report from the technical team at Linode and it seems they are aware of the connectivity issues with Mumbai Datacenter and doing their bit to resolve it.
The networking issue is there since their launch in India and I don't know if this will be resolved anytime soon as it has been reported a lot of months ago and there hasn't been any improvement so far.
Here is the latest response from the team:
Thank you for that MTR. I will make sure to pass this on to our network engineers. We are currently working on improving connectivity in Mumbai by partnering with additional ISPs. In this case, the issue arises on an ISP several hops outside of our network, and our ability to improve connectivity to wordpress.org from our Mumbai datacenter is extremely limited.
As you can see in the MTR you provided, latency and especially packet loss increase sharply around hops 11-12, and this persists until the traffic eventually reaches Wordpresses' website.
One of our ISPs has responded, and cannot reproduce this issue on their end. We are actively reaching out to our upstream providers in an attempt to improve connectivity. We will keep you updated with developments on that front as they become available to us.
I already have a USA Based Linode server in my account which utilizes the same server setup and a WordPress website. It has been around 2-3 months since I deployed the USA based Linode and I do weekly updates. The plugins, themes, and WordPress Core updates are done in seconds on the website hosted at USA based DC. I haven't faced any similar issue at this particular server so far.
The problem with networking seems to be with Mumbai DC only. I hope this issue will be taken at a high priority and will be resolved soon :)
I have multiple websites hosted on Linode. Plugins and core update is pending on all websites. I turned frustrated and I have started to move out to DigitalOcean. Surprisingly, it took me barely 10 minutes to move 3 websites and updated everything properly.
Same issue I have faced multiple times earlier too. This time I could not be as patient as before.
I don't know how long they gonna take to resolve, I have made my mind of migrating out completely in next week.